Articulating Architectural Design through Computational Media
by
Mark John Sich
Bachelor of Science, Building Science 1992
Bachelor of Architecture 1993
Rensselaer Polytechnic Institute
Submitted to the Department of Architecture
in partial fulfillment of the requirement for the degree of
Master of Science in Architectural Studies
at the
Massachusetts Institute of Technology
June 1997
@ 1997 Mark J. Sich All Rights Reserved.
The author hereby grants to MIT permission to reproduce and to distribute publicly
paper and electronic copies of this thesis document in whole or in part.
Author:
Mark
Sich
Thesis Supervisor: William Mitchell
L(
/
.1
Department of Architecture, May 9, 1997
Prof. of Architecture & Media Art and Sciences
-,I
Chairman, Dept. Committee on Graduate Students:Roy Strickland Associate Professor of Architecture
OF TECHNOLoGY
JUN 2 01997
LIBRARIES
00. 01
ARCHNES3
Reader page:
Reader:
Thom Mayne
Principal of Morphosis, Santa Monica
Professor of Architecture, SC-arc & UCLA
Reader:
Takehiko Naqakura
Assistant Professor of Design and Computation
00. 02
Articulating Architectural Design through Computational Media
Mark John Sich
Bachelor of Science, Building Science 1992
Bachelor of Architecture 1993
Rensselaer Polytechnic Institute
Submitted to the Department of Architecture on May 9, 1997 in partial fulfillment of the requirement for
the degree of Master of Science in Architectural Studies at the Massachusetts Institute of Technology
Abstract
This thesis proposes the concept that computational tools can merge two high order
cognitive operations for architectural designers. The first cognitive operation being the internalization of displacement (as defined by Kant and Piaget). The second being the construction of a mental Model (i.e. one that represents a tended external reality). The existing computational tools allow for a new external representation that has a relationship with the internal representation in the designers mind. The conceptualization of complex systems is in a
direct relationship with the designers ability to visualize infopro (information processing) paradigms. The designer must be afforded the opportunity to undergo the cyclic process of conjecture, reading and evaluation.
How can an architects sensibility to physicality be integrated into computationally
based modeling and representation. The larger contextual questions are; What are the inherent differences between actual physical models and computational models? What is lost and
what is gained, and can they be reconciled with each other? What are the cognitive operations that are assisted through the use of these new computational models? As these questions are broken down, it becomes apparent that the difficulty of re-presenting the computational model back to the designer (and others) must be investigated. Interaction in virtual
space is not new, but the investigation of an application using an architectural setting has
not yet been explored.
I propose the creation of an immersive tool that will allow for a phenomenological
creation of a virtual model. The application will examine the interaction that occurs between
"reality" and proposed reality. The addition of an abstractive navigation system will help facilitate the conceptualization of a congruent infopro model. When the program is completed
(first 1/2 of the semester), different design approaches will be tested by different end users
with different physical interaction models. The output that is generated will lead to the proposal a new typology of spatial forms. The proposed implementation will incorporate the
presentation model, but only for reflection, the primary purpose is a computationally centric
design process.
Thesis Supervisor: William Mitchell Prof. of Architecture & Media Arts and Sciences
00. 03
Acknowledgements:
I would like to extend my sincerest thanks to my Advisor, Bill
Mitchell. Your insights showed me the path.
To my parents, Eugene and Jutta, Your unconditional love and
sacrifices gave me the means and the vision.
To Daniel J. Brick, your support and friendship that you
unselfishly gave, pulled me thru.
To Edgar Ngwenya, your skills and patient teaching, gave me
the method.
To Thom Mayne, your mentoring gave me the chance to prove
myself.
Last but definitely not least, to Andrea M. Stancin, your undying
support and love carried me through this all. YOURS.
To all my friends at MIT, I am truly blessed to encounter so many
wonderful people. I can not thank you all enough for everything
that you have done. You will be in my thoughts
forever
00. 04
Table of Contents
Title Page
reader page
02
Abstract
03
Acknowledgments
04
Chapter 1
Introduction, overview
Directions
08
13
Chapter 2
Design Biases
Chapter 3
Cognition
Physical Vision
Physical Energy
The Eye
Optic Nerve
Visual Cortex
Implications
19
21
21
22
25
25
27
Chapter 4
Open Inventor
Nodes/Fields
VRML
Scheme
IVY
29
30
31
32
34
Chapter 5
Media Manipulation
Movement and Navigation in PHEMENOarch
Manipulation in PHEMENOarch
38
40
00. 05
Table of Contents ||
Chapter 6
Test Methods/Features
SoTabBoxManip
SoTransformBoxManip
SoJackManip
SoHandleBoxManip
SoTrackBallManip
SoTextu re
SoPointLight
Testing Methods Proper
42
43
44
45
46
47
48
48
49
Chapter 7
Physical Manipulation/Interaction
The Computational Drafting Board
Semi-Immersion
The Brunelleschi Illusion
The "Cave"
The DIGIPANOMETER
Virtual Reality
52
54
56
56
60
63
65
Chapter 8
The Test Experaments
Reflections on Abstractions
Methodology: Experament (the Blue phase)
Methodology: Case Study
Edge Boundaries
Light
Transitional Boundaries
Textures (trouble with)
68
68
69
75
81
83
84
85
Chapter 9
Conclusions
86
00. 06
Table of Contents Ill
Code
89
Appendix A
To set up the environment for PHEMENOarch
101
Appendix B
To use PHEMENOarch
103
Appendix C
.iv Files and Alterations
106
Bibliography
115
CO.
07
S I
m
INTRODUCTION
Overview:
Architecture is a means of communication.
It solidifies ideologies, it takes a view. The purpose of design is to slide behind the facade of
appearance and to explore the underlying truths
we do not normally see, Pursuing idealized and
idiosyncratic realities inherent in every project
and form. It is my belief that the computer can be
instrumental in these endeavors.
This thesis is aimed at a discourse
into the tacit relationships between the exploration
of the imagination, with all its inherent ambiguity,
k 1\
0 I .0 I
Description:
Computer
rendering of
Vladimir Tatlin's
1920
the Monument to
the Third
International
FIGURE
Architecture is a
Political act and a
means of
communication
and the manner in which we re-represent our
ideas back to ourselves for re-evaluation.
The context of Architecture is shifting. We are
entering a new era in which our specialized languages of space and form can now be visualized and experienced by all and not just the few
who can read and interpret plans, sections and
models. For this empathic visualization is what
OI. 08
I~I
we are all about, we (architects) live in a virtual
world of ideas. This world must be conveyed in
such a precise manner that it can be brought into
reality by hands other than our own. The newest
and one of the most powerful tools, the computer,
is emerging as a forerunner to help us with our
almost insurmountable tasks. The implementation
of the tool must be tacitly understood and critically applied. In my opinion, the computer is not
being used correctly in most of its architectural
applications. It is being used either as a simple
tool of mass production or solely as a simple tool
of presentation. Almost no one is using it as a
design tool. The great power of the computer
does not lie in the fact that it can perform 30 million recursive operations every second, but rather
FIGURE
0 I .02
Description:
Salick Health Care
Computer driven
design of facade
and entrance
it lies in the fact that the computer affords us a
reflection into our own minds. What are the
issues in this redefinition of architecture and architectural design? These are the areas I endeavor
to explore.
The scope of this thesis is to look at the
inherent strengths and weaknesses of designing
Architecture with in virtual environments. I argue
that what I call "traditional" architecture has been
constrained in regards to space and form creation, because of the inherent difficulties of
abstracting oneself into an unfamiliar frame work.
As a result, one found the necessity to rationalize
I1.
09
~1
the environment around you and create a formal
description of precedent, in order to begin to
gain an understanding of the design that one is
developing. In order to get back to the main
issue of space creating, i.e. creating environments in which people live and work and play, I
propose the using of the computer as a design
tool of immersion. The conception and implementation of Architectural design has been closely
aligned with the manner and technology with
which it has been represented. For the first time
in Architectural history, one can "build" the design
proposal and then evaluate it from within, allowing for an infinite possibility of changes and iterations from a phenomenological perspective. One
can achieve this and be economically viable as
FIGURE
0 I .03
Description:
Interior of
Actual Salick
Health Care
Entrance
well. Architecture relies on an implication of
space and it's experiential qualities rely on
motion through built space.
The implementation of the computer as the
primary tools of design is an idea which iscoming of age, schools and universities around the
world are taking the first tentative steps into this
"new" realm. One example of this is MIT's virtual
studios, but even here on the cutting edge, the
direction of CAD development in the realm of
architecture has been, on the whole, towards the
later stages of design development. The machine
has acquiesced into its role as a valuator of ideas
OI.
IO
or a communicator of established ideas across
vast distances. This is obviously a very worthwhile
endeavor, however I stand with the belief that the
computer is not being pushed to its fullest extent.
Erich Mendelsohn sought to integrate the
tensions of structure and form into a controlling
image, he utilized a romanticized notion of the
sketch as a tool of envision enabling. Jackson
Polic used his psychoanalytic drawings. Gestalt
psychology explored notions of subject/objectification. Felix Candela stated that, "a true work of
art needs no justification by esoteric theories and
is not necessarily the result of a process of conscious reasoning." This line of thought leads to
FIGURE 0 I .04
Description:
Erich
Mendelsohn's
Einstein Tower
Potsdam, 1919
my investigation the hypothesis that abstraction
alone does not afford the designer the opportunity to fully undergo the cyclic process of conjecture, reading and evaluation. Using the limited
scope of Visualization (there are other factors to
architecture and its analysis), This thesis proposes
the concept of designing within an immersive
environment.
The utilization of these computation tools reevaluates the phenomenological approach within
post-modern architecture. For this very reason
01. 11
qi
Immersion and engagement are ultimately necessary for the stimulation of a virtual architectural
experience and it's evaluation.
The issues of computer computation are on the
forefront due to our abidance with the concepts
of modernity. (What is new is better than what is
old.) I take issue with this mentality, in 20 years,
not a soul will be intrigued with the concept of
the possibility of using the computer in the
design process, it will have become a reality. This
not to say that the role of models and sketches
will necessarily be diminished, but there will be
role shifting and the new affordances will dictate
how and why a medium will be chosen.
FIGURE
0 I .05
Description:
Experiential Image
of Santiago
Calatrava's '92
expo pavillion
The computer allows us another extension of our
minds beyond the boundaries of the body, This
ability has yet to be explored, and I suppose that
the reason is due the underlying fact that the computer is acting as a mirror back into our own
minds and souls. Hence, I believe that it will
never be completely understood. The vigilance of
computation must emphasize that the machine
should not be used to quantify our existence and
dictate our hands. This tool is bigger than all that.
It must let us explore.
01.
12
Directions, the fork in the road.
One of the decisions that became prominent during the development of this thesis was the course
that the development of the application should
take.
The way that I saw the development of the application was in one of two directions. The first
j L_
development approach was to create a tool that
could stand as "THE" design development tool for
architects. One would have to take advantage of
the all levels of Abstraction (i.e. Plans, Sections,
elevations etc...) plus have all the phenomenological feedback of an immersive environment. The
interaction between these two methods of visualizing design space would have to be very carefully
investigated from an architects perspective.
Initially my design strategy for the application
was directed in this manner. I quickly discovered
the overwhelming complexity of such a task. The
physical development of the tool became but one
FIGURE 0 I .06
Description:
The construction of
a perspective from
the abstractions of
plan and elevations
of the challenges. The refinement required to integrate two separate information processing systems seemlessly became a tiring goal. The deeper that I progressed in my development, the more
it seemed that I was shifting away from the main
focus of the thesis, the investigation of designing
within virtual space. At this time I had also noted
01.
13
that many of the commercial CAD packages
would perform most of these tasks, to a greater or
lesser extent, better than what I and this avenue
of development could produce.
The second manner of application development,
and for this thesis, the more fruitful, was to focus
o-xvnx
73
on the development of INFOPRO models, in a
cognitive sense. The goal of the tool making
became the development of a phenomenological
3D space manipulator that was so specialized
that the designer could interact with the virtual
space until the model and the designers internal
representation became one. This goal lends its
self to testing quite easily. If there is enough evi-
FIGURE
0 I .07
Description:
Example of
INFOPRO
information
processing
dence that the level of error between intent and
model is minimal enough, the results can be used
to demonstrate other meta-cognitive operations
that the designer may be embarking upon subconsciously.
01.
14
The resultant application is called;
PHEMENOarch (pheenomeo-arch)
The name derived from two parts.
The first part comes from the greek "philos",
meaning the study of phenomena. The name may
also be interpreted as relating to PHEMENOlogy,
the study of phenomena, which are the objects of
knowledge, the only form of reality (1860), the
view that all things, including human beings, consists simply of the aggregate of their observable
sensory qualities (1865).
This is paired with the principals of ARCHitecture
and architectural thought. Architecture, in my
opinion, defies definition to a certain extent. It
means so many slightly different ideals to different
individuals. I am defining architecture, in this
instance, as being the process of space creation
and evaluation.
01.
15
-
A -
-
DESIGN biases,
Everyone is a prisoner of his own experiences.
No one can eliminate prejudices- just recognize
them.
-Edward R. Murrow
As with any investigation, the research will
always be imperfect due to the imperfection of
the researcher. There are inherent biases that all
of us have developed. In order for the reader to
better understand where the thesis is originating, I
believe that I must disclose some of my personal
design methodologies, so that the readers of this
thesis can begin to untangle my bias to a certain
extent.
FIGURE
02.0 I
Description:
Leonardo da Vinci
extension of the
body
"the designer is always faced with a decision of
intention and not merely a deterministic process",
-Alan Colquhoun, typology and
design method
In thinking about design processes, it is difficult to
isolate episodes in a design process. One can
investigate and observe a series of alterations of
02.
16
I
a singular design move using different methods.
For instance, one can see medium explorations
quite readily, when does a particular designer go
from model to sketch to section? One can also
note geometric vs. topographical paradigm shifts.
However, I argue that the designer does these
shifts not in relation to some predestined set of
rules but rather by using his/her intuition as to
when to shift paradigms. A good designer will
never travel the path of creation the same way
twice. The requirements of each design bring out
a necessity to explore certain criteria with a specific design hierarchy.
The methodology for this investigation can be
charted into 3 distinct phases. The first is conjecture, the true act of creation. One asks the question "what if?" The second stage is a re-reading
FIGURE
02.02
Description:
The Cycle of
Design thought
and Investigation
of the form or decision. "what is it that I have just
created?" The last phase is one of judgment.
"Does this move make sense? How can it function
more effectively? etc...." Once this evaluation is
complete, the designer is coerced into the cycle
again by trying to answer the questions that the
judgment phase left unanswered. One reflects at
a meta-cognitive level about operational-level
moves and strategies, a realization after the fact.
I also believe that Helmholtz's constructivist perspective theory comes into fruition during design.
The primary focus of constructivist theory is that
02.
17
qi
perceiving and thinking are active goal-oriented
processes. One idea leads the mind and the perception to attempt a certain range of design
strategies much like the eye performs saccades, a
scanning over an image in order to understand
the gestalt, and Schemata, which proposes
where the eyes should look next in order to more
rigorously investigate the object/image.
Using Don Norman's concept of affordances.
What does the tools feedback entail, encompass
or mean? What does it allow you to do?
A design tool for designers and architects must
afford the designer the opportunity to under go
these cyclic phases. Any tool that elevates the
architects awareness of the full implications of the
design moves that he/she are doing, will ultimately increase the effectiveness of the architect.
There by creating a more robust and fleshed out
architecture.
FIGURE
02.03
Description:
Eye Saccades
over
bust of Queen
Nefertiti
02.
18
COGNITION
The first steps towards an immersive tool must be
tempered with the understanding of how we cognitively understand the world around us.
Xf
In this chapter, I will discuss how the mind
sees and understands visual input. The last few
years have seen a marked increase in the science of human cognition. Human cognition is
described as the branch of psychology that investigates memory, perception and thinking. I conject
that with it's understanding one ultimately gains a
better understanding of human psychology and
human design processes. This understanding is
extremely relevant, especially when we investigate design inception.
To begin, Human beings have been gifted
with a myriad of senses. We have the senses of
audition, taste, touch, smell and vision. It is a full
combination of all these s stimulated senses which
informs us of our placement and environment. It is
this culmination that makes the implementation of
FIGURE
03.0 I
Description:
Descartes's Eye
La Dioptrique 1637
OX eye experiment
demonstrating image
inversion.
immersion in a virtual setting so difficult. To truly,
be "immersed" all these senses must be accounted
for and stimulated in real time. However, the status of computational power is not yet ready to
03. 19
adequately tackle such an endeavor. So as a
result, of the different senses to investigate, the
faculty of vision is the most influential in the dialogue of design and perception. This is where we
will begin.
"Vision... is the big window", Robert L. Solso,
1994
FIGURE
03.02
Description:
The eye, often
thought of as the
window to the soul
03. 20
PHYSICAL VISION
Vision is an analytic operation of "divide
and conquer". The basic path for visual information is
physical energy->eye->visual cortex->associative
cortex
These are described in further detail.
Physical Energy (Light):
The eternal world is alive with different
electromagnetic radiation. One form of which is
visual light. This Electromagnetic radiation is created by electrically charged particles. This radiation has a definable physical characteristics,
FIGURE
03.03
Description:
Sir Isaac Newton
using a prism to
split sunlight into
spectral colors
these are described in wavelengths. These waves
are then reflected and/or absorbed by differing
degrees when they collide with objects of varying
materials. These There are two seemingly opposing theories of light, one theory states that light is
a stream of particles. This theory was first proposed by Sir Issac Newton (1642-1727). The
other theory of light, by Christiaan Huygens
(1629-1695), was that light is a pulse or a
wave, that it had no physical form. Without getting in to contemporary physics, they are both
03. 21
right. Light behaves as both a particle and as a
wave, but not at the same time. The computational modeling of this phenomenon will be discussed
in a latter chapter.
X
1"
FW.
W
AM
A, .-,yWW
The Eye:
The eye is the organ which is able to
detect the differing wavelengths that are reflected
by the environment. Commonly, humans perceive
electromagnetic wavelengths that range between
375 nanometers and 780 nanometers. The shorter wavelengths are seen as violet or blue, and
the longer wavelengths are seen as reds. Greens
and yellows fall between these values. The amplitude of the light also effects the perceived bright-
W,*ocv*0
*rt
'V1V
FIGURE
O''*
Ri
03.04
Description:
wavelenghts in the
electromagnetic
spectrum
ness of the light, a smaller amplitude will be registered as a dimmer light.
Light enters the eye through the pupil and
is focused. The cornea is found on the surface of
the eye and is followed the iris and then the lens.
The lens does not actually bend the light, the
aqueous humor (fluid) with in the cornea performs
this task. The ciliary muscle deforms the lens in
order to focus on objects. A closer object is in
focus when the ciliary muscle is relaxed there by
reducing the curvature of the lens. Consequently
objects further away focused by tightening the
muscle forcing the lens into a deeper convex configuration. The mass of the eye is made up of
03. 22
-U
Vitreous humor, a dense liquid that maintains the
form of the eye.
Light then strikes the back of the eye. This
area is called the retina (from the Latin "rete" translated as "net"). The receptor cells, of which there
are two types, are sensitive to electromagnetic
energy. Rods are receptor cells which are more
sensitive to grays and lower intensity light. Cones
are receptor cells which specialize in color differentiation. These cells translate light energy into
signals that are passed to the bipolar cells and
then to the ganglion cells (collectors), which in
turn pass information to the optic cortex.
With in the eye proper, there are about
FIGURE
03.05
Description:
Cones of vision,
relating foveal,
parafoveal and
peripheral vision
125,000,000 rods and 6,500,000 cones. The
distribution of rods and cones across the retina is
not uniform. There is a 2mm. indentation opposite
of the pupil called the Fovea. The Fovea or
Macula Lutea (yellow spot) is densely packed
with Cones and no Rods. It is this distribution
which accounts for the reason that a truly focused
area that one can see is only 1-2 degrees. This
area of focal vision is called Fovial Vision.
It might be said that by moving from the center of
the human retina to its periphery we travel back
in evolutionary time; from the most highly organized structure to a primitive eye, which does little
more than detect movements of shadows. The
03. 23
very edge of the human retina... gives primitive
unconscious vision;and directs the highly developed foveal region to where it islikely to be
needed for its high acuity.
-Richard L. Gregory
The eye can detect or perceive visual phenomenon with in a field of 180 degrees horizontal to approximately 130 degrees vertical. Fovial
k
/ f~- 7
Ovtw~cnf
vision is as mentioned about 2 degrees, however
parafovea, the area that is perceived as focused
although less defined, is up to 30 degrees.
These limitations are the cause of our necessity
for constantly scanning our eyes across an image
to gain an impression of the full object.
The understanding of these fundamental principals
of human vision are essential for integrating a virtual design environment. The interesting computa-
FIGURE
03.06
Description:
Visual Fields
limits of monocular
and binocular
vision.
note:
distortion of symmetry caused by
occlusions from the
nose.
tional advantage is that one must only render a
small area very accurately and the rest can be
approximate, if one can project the virtual environs exactly where the eye looks. This is an
intriguing concept but not in the scope of this the-
sis.
03. 24
Optic nerve
The eye passed its information on to the
optic nerve. These nerves (one for each eye) are
merged at the optic chasm. The information from
the eyes are split such that the visual fields (not
what information are from which eye) are passed
_,ebaV
to the opposite visual cortex. I.E. the right half of
your perspective viewpoint is transferred to the left
Optic
side of your brain, even though the actual information is coming from two distinct eyes.
Visual Cortex
The visual Cortex receives the visual infor-
cortex
mation, responds to specific visual stimuli. This
infers that images are broken down into a specific set of stimulated neurons. If the image changes
another batch of neurons will be stimulated. Each
cell seems to have a "particular shape of stimulus
and to one particular orientation" David Hubel
FIGURE
03.07
Description:
Neural Pathways
visual fields and
their associated
brain hemispheres
(1963)
David Hubel and Torstin Wiesel shared the 1981
Nobel prize for physiology and medicine.
The implication is that we do not see a
holistic object (which was once thought) but
rather we place value upon certain stimuli and
reassemble an image in our minds.
03. 25
This is backed up by an intriguing study by Levy,
Trevarthen and Sperry in 1972. They studied
commissurotomized patients, these are people
who have had their corpus callosum cut. Corpus
callosum is the connective tissue that allows the
two halves of the brain to talk with each other, if
this tissue is severed the two sides of the brain
can not communicate. When these patients were
wj
/
shown a chimeric face (in this case the left half a
woman and the right half a man composted
together), an interesting pattern emerged.
When the patients were asked to describe
the face verbally, they picked the man to
describe. This fostered the understanding that the
left hemisphere the side of the brain that computes the right side of an image, is associated
with verbal information processes.
r~p
These same patients were then asked to
pick the face that they were just shown from a
selection of whole faces. They chose the
woman's face. This seems to imply that visual and
or pictorial processing is done through the right
side of the brain.
Harris 1978 p4 6 3
FIGURE
03.08
Description:
Chimeric face
used with commissurotomized
patients
The left hemisphere operates a more logical, analytic, computer-like fashion, analyzing
stimilutus information input sequentially, abstracting out the relevant details to which it attaches
03. 26
verbal labels:
The right hemisphere is primarily a synthesizer, more concerned with the over all stimulus
configuration, and organizes and processes information in terms of gestalts or wholes.
IMPLICATIONS
This cognitive information would seem to imply
that designers and Architects are more deeply
involved with right hemisphere brain processing
than left. It can also be concluded that this hemisphere of the brain should be stimulated and/or
fostered to a greater extent during design charretts.
This cognitive understanding also points to the
fact that almost all software that is developed is
FIGURE
03.09
Description:
AUTOCAD
interface
note:
The left side
biased menu
system and the text
based command
line. Both are not
conducive to Right
brain thinking/
creativity
incorrectly set up for architectural thinking processes. According to the afore mentioned studies, the
menu bars of software should be placed on the
right hand side of the screen in order to more
effectively stimulate the left hemisphere of the
brain, the side which is concerned with labeling
and analysis. Along with this directed consciousness concern, the main work space of a program
(i.e.. the view space/interaction zone) should be
placed on the left side of the screen. This placement would stimulate the right side of the brain,
03. 27
I-
I
IIII
the side concerned with the understanding of the
whole picture.
Ifone were to examine the popular softwares of
our time, one would discover that they are all
laid out in exactly the opposite way. They are all
reversed. And some of the most pervasive CAD
software of our time, like Autocad, have yet
another difficulty. This difficulty stems from the fact
that they are text based interfaces (for the most
part) which are left hemisphere intensive but the
work of 3 dimensional modeling and creation is
more of a gestalt thought process requiring right
hemisphere thinking. The two methodologies are
counter intuitive and counter productive for the
FIGURE
03. I 0
Description:
PHEMENOarch
Interface
most part.
In the software, PHEMENOarch, that I have created, the emphasis of the interaction with the
geometric forms and their generation stems from
a right hemisphere stimulation approach. The
designer is rarely removed from the designing
process by being forced to undergo change of
thinking through a menu system. The main menu
items that one is required to access, are displayed in the right side of the screen in an
attempt to minimize concentration shift paradigms.
03. 28
A -
-
PHEMONOarch is scripted/written with Open
Invento. It makes sence to give an overview of
the affordances of this language when couppled
,ve
with Scheme .
Open Inventor
Open Inventor is an object-oriented 3D
graphics toolkit. Each "object" on the screen is
represented as an "object" in memory. For those
of you fortunate enough to have taken 6.001,
this is like the object-oriented "scheme-builder"
FIGURE
game.
The Architecture
04.0
Description:
of Open
Inventor
Each object is stored in the scene database, SoDB, which stores such information as the
name of each part in the current scene. The
SoDB then holds the objects in tree like data
structures. This fundamental data structure in
Open Inventor is called the scene graph. A scene
graph is a directed acyclic graph. Most of the
operations in Open Inventor involve applying
actions on to the scene graph proper. These
actions traverse this tree from left to right, and
from the top down.
For example, when Open
Inventor renders a scene, it does so by applying
a render action to the scene graph. When a program wants to find a certain node in the scene
04. 29
I
graph, it does so by applying a search action to
the scene graph. It was this methodology that
was used in constructing the alterable selection
nodes in PHEMENOarch. The scene graph paradigm that Open Inventor uses, has been shown
Group
to be a successful one. This paradigm has been
adopted in many current and forthcoming 3D
graphics standards.
*
Separator
.
Transform
Most of the programmatic interaction involves
nodes in the scene graph with Open Inventor .
Nodes often contain data elements known as
Property
Transform
Shape
fields. These Fields encapsulate information that
Open Inventor uses to perform certain actions,
like rendering a scene or object.
Nodes
FIGURE
04.02
Description:
Primitive example
of a Scene Graph
and it Path (dark
line)
A node is an object which stores some
usable piece of information. Each node has an
associated data type, or class, which defines the
operations, or methods, which can be performed
on that node and the data which that node can
store. All nodes are derived from the base type
SoBase. This means they are specializations of
this data type; a node can do everything an
object of type SoBase could do, and more.
Fields
Nodes usually contain data elements within
04. 30
fields. Each field within a node isalso a complex
object, rather than a basic data type. The three
main reasons for doing this are: All field types
have consistent set and get functions. Hence with
the use of setValue and getValue, one can
retrieve values stored in a field of all field types
because they are consistent. Also a node can tell
when the value of a field stored within it has
changed, and can automatically notify the application. The last reason is that because Fields can
be connected together in chains, the manner in
which Open Inventor screens it value is by automatic propagation down the chain. This mechanism is known as notification and can be used to
reinform the application of actions to be taken
and/or have been taken.
FIGURE
04.03
Description:
Property-Node
Classes in Open
Inventor
VRML 2.0
VRML 2.0 is the current standard for displaying 3D graphics on the World Wide Web.
Many spectacular example abound, For more
indepth reading investigate Daniel J. Brick's thesis
on Virtual site interactions. VRML is based on the
Open Inventor ".iv " file format with a few added
features. VRML carries over the concepts of the
scene graph, nodes with "fields", engines, and
sensors. Most VRML browsers are implemented as
Netscape plugins.
There are explicit differences between VRML and
04. 31
Open Inventor. Fore instance, children are stored
explicitly in "children" field of group nodes. Fanouts, where a single engine create multiple
nodes, are allowed, but fan-ins, where multiple
fields feed a single engine, are not.
SCHEME
Scheme is a statically scoped and properly tailrecursive dialect of the Lisp programming language invented by Guy Lewis Steele Jr. and
Gerald Jay Sussman. The design intent was to
have an "exceptionally clear and simple semantics and few different ways to form expressions".
A wide variety of useful programming paradigms
( functional, imperative, and message passing
styles) are made readily convenient in Scheme.
FIGURE
04.04
Description:
Scheme at MIT
6.001 badge of
Honor
Scheme was one of the first programming languages to incorporate first class procedures as in
the lambda calculus. Scheme was the first major
dialect of Lisp to distinguish procedures from
lambda expressions and symbols. By relying
entirely on procedure calls to express iteration,
Scheme emphasized the fact that tail-recursive
procedure calls are essentially goto's that pass
arguments. Scheme was the first widely used programming language to embrace first class escape
procedures, from which all previously known
sequential control structures can be synthesized.
More recently, building upon the design of gener-
04. 32
ic arithmetic in Common Lisp, Scheme introduced
the concept of exact and inexact numbers.
Scheme recently became the first programming
language to support hygienic macros, which permit the syntax of a block-structured language to
be extended reliably.
-Kenneth B. Russell
Due to the fact that Scheme is an interpreted language one can interactively edit Inventor programs and scene graphs. One can even include
Scheme code in Inventor scene graphs.
Callbacks, for instance, can execute Scheme
code rather than call a C function. These callbacks can then be interactively modified at run
time.
04. 33
|vy
If you are running the application PHEMENOarch
on Athena, one will have to set up Scheme with
the base Inventor code. Ivy is a Scheme binding
for Open Inventor.
http://www-swiss.ai.mit.edu/scheme-home.html
Ivy works by providing a consistent interface to
the member functions and class variables of the
Open Inventor classes. The underlying C++ functions are called from the Scheme backend. From
the user's point of view, it is a simple syntactical
change.
04. 34
-
A -
m
Media Manipulation
How manipulation should behave in
PHEMENOarch
The task of manipulating objects in 3-dimensional
space is a difficult one at best. The human mind
is used to operating in a 4 dimensional world
space (i.e. the 3 axis of x,yz plus the dimension
of time) but we rarely need to THINK in 3 dimensions, let alone 4. Attempt to solve any of the
metal linked ring puzzles or Rubix cube at your
local mall and you see exactly what I mean. The
solutions are a simple game of geometric manipulations, however when they are coupled with the
element of movement (a 4th dimensional mental
manipulation, i.e. having to pass this ring through
that slot, while twisting the metal angle into position etc...), these solutions become extremely difficult to conceptualize and even harder to realize.
The difficulty of visualizing true 3 dimensional
FIGURE
05.01I
Description:
Interlocking
Metal Ring
Puzzle
space is even translated to the experienced
designer/architect. Rarely, if ever, does one find
an architect that designs directly in axon or
model, the usual methodology is to break 3
dimensional space down into 2 dimensional
05. 35
abstractions of plans and sections. These abstractions allow the designer to formulate relationships
and a parti. This parti then develops into an ever
increasing array of spatial abstractions (sections,
details, built models etc...), as the designer is
able to slowly internalize the implications of the
space and forms that they are creating. I would
not attempt to imply that this methodology of
decomposing a large problem into smaller more
manageable abstractions should be abandoned
or dismissed. Rather, I see it as a need to be
able to allow the designer more tools at his/her
disposal in order to better understand what the
proposed design really is. Inthis particular thesis,
the concentration lies in the need to relate to the
designer, in an intuitive manner, just how forms
FIGURE
05.02
Description:
simple examples
of the difficulty
of visualizing
spatial complexities
and spaces can be constructed in virtual space.
The manipulation of the virtual tools must be crafted as to not break the cognitive infopro model of
immersion. The tools must allow for the user to be
able to manipulate, change and create forms and
spaces without ever having to look at a menu or
other left hemisphere abstraction models like text
or menus.
The manipulation tools must be intuitive to use.
05. 36
This philosophy towards interface and interface
Cursor movement
design required a reevaluation of all the commercial Modeling packages. Through this research of
existing modelers, the realization dawned that
this methodology to spatial creation had never
Cursor kcation
been implemented before.
The packages that least exemplified the concept
of intuitive manipulation of spatial relationships
and forms were those that required a cognitive
effort to switching between creation, manipulation
and evaluation nodes. The most notable of which
was Autocad, in which to even move an object
FIGURE
05.03
Description:
Autocad
VPOINT
viewpoint
manipulation and
evaluation tool
vertically required either typing in the coordinates
or changing the ucs (user coordinate system) to
another reference plane (x,y or z), shifting to
"home" in order to even see the form in elevation/plan, changing modes to move the form
(requiring at least 3 mouse clicks), moving the
form and finally evaluating the form with
"VPOINT" which does not even give you a preview of the viewpoint until you have accepted it.
The next level of intuitive use, would be the
Alias/Wavefront and Form*Z packages. Alias
allowed for the designer to evaluate the created
forms easily and usually in smooth realtime. The
FIGURE
05.04
Description:
Alias
viewpoint
manipulation and
modeling interface
form model creation paradigms (especially if you
are dealing with splines) are very intuitive in that
you can see your changes on the screen automatically. The menu system and other manipulators
05. 37
leave a lot to be desired however. It is very easy
to become lost or confused with the hap-hazard
methodology of menu tools. Form eZ allows for
an even more intuitive object/form creation, but
its evaluation system is not as advanced as Alias.
This last difficulty may stem from the power of the
platforms that both software packages are written
to use, Alias for high power/ high end SGI
machines and Form*Z for lower end Macintosh
and Windows NT machines.
PHEMENOarch does not pretend to have the
FIGURE
05.05
Description:
screenshot of
PHEMENOarch
created
Geometry
modeling capabilities of even the lowest level of
commercial modeling softwares, the interface,
however, does bear some scrutiny.
Movement and Navigation in
PHEMENOarch
When PHEMENOarch is married with a Virtual
Reality system, the navigation is literally as simple
as swiveling your head around to see the objects
for evaluation. This exactly mimics what we do in
the physical world. This "hands-off" visualization
technique allows for the designer to concentrate
FIGURE
05.06
Description:
Object is selected
SoTabBoxManip
placed
on the tasks of conjecture, reading and judgment
with in the act of creation.
05. 38
When PHEMENOarch is placed on a regular
display system (many of which will be discussed
in a later chapter), the navigation is still intuitive.
The viewer is constrained such that the designer
move about the space in an analogous manner
to walking. One may move forward or back in a
horizon plane, turn corners or stand still and look
about. The designer is also afforded the ability to
elevate his/her viewpoint by consciously requesting the elevation change with a click of the
mouse. There is no freeform "flying", which is a
characteristic inherent in the other browsers that
FIGURE
05.07
Description:
SoTabBoxManip
unidirectionaly
scaling cube
have been investigated. To see an example of
this "flying" iust look at the proliferation of bad
animations that have recently become abundant
in which one shoots through the latest developer
community at a height of 150 ft, then dives into
an open window (or through a wall) and zooms
past the kitchen. This may be flashy but it has
very little to do with human inhabitation and
viewpoint. PHEMENOarch allows for the operator to maneuver as if he/she were in a real spatial environment. One does not suffer the loss of
placement within a space. Every move is contiguFIGURE
ous. This is not to say that one can not get lost in
a space, but the proclivity for it is decreased.
05.08
Description:
SoTransformBox
Manip toggled
05. 39
Manipulation in PHEMENOarch
The careful consideration of movement and
manipulation had to be investigated.
PHEMENOarch's selection system is straight forward, one places the cursor over the form to be
manipulated and select. A manipulator appears
attached to the form that has been selected. The
manipulators can then be utilized or toggled to
the next manipulator, which can then be used or
FIGURE
05.09
Description:
Geometry rotated
through z axis
toggled yet again. All your choices are immediately at the disposal of the designer. There is no
need to go to a menu bar to manipulate scale,
rotation and/or movement parameters. This
methodology retains the immersion of the designer such that it feels as if one is able to move the
walls out, the ceiling down, etc.
Open Inventor offers with in its toolset powerful
selection manipulators. The strength of these
"manips" are that they have an iconography built
in that is readily intuitive for the designer to internalize and use. To rotate a form, for instance, is
FIGURE
05. I 0
Description:
SoTrackBallManip
selected
a simple act of grabbing an edge and sliding it
in the desired rotation direction, the SoDB is
instantaneously up dated such that the screen displays the changes interactively. One can see the
05. 40
updates as one works, in realtime. The investigation of the implications of these individual manipulators are discussed in chapter 6.
There was also one other concern that needed to
be addressed in a serious manner. The method of
primitive form creation. In the earlier versions of
the program, new geometries were always created at the origin. This methodology was acceptable for small, compact models. However as the
FIGURE
05. I I
Description:
Cube rotated in
free space
models increased in size and complexity, the
designer would have to return to the origin in
order to "make supplies", and then they would
have to transport them to the site where they
would be used. This was analogous to a bricklayer going to the palate to get more bricks. The
methodology was not very conducive to a seamless creative environment. The solution was to
have the SoCube create the primitives in relationship to the objects that are currently being manipulated (otherwise if no other objects were selected, it would just create the primitives at the origin). With this methodology, the designer could
continue creating and not have to worry about
resource management.
FIGURE
05. I 2
Description:
SojackManip
selected and
geometry is moved
along internal
forms axis
05. 41
m
Testing methods;
In order to understand some of the testing methods, a few of the applications various features
and versions must be explained.
Several versions of the Application
PHEMENOarch were created, in order to test the
effectiveness of differing abstractive barriers and
functions. Especially what functions were conducive to differing designing styles and attitudes
toward the design process.
FIGURE
06.0 I
Description:
Phemenoarch
created form
The most noteworthy are the differing manipulation tools. These were tested separately to understand their affordances, before they were combined to provide a fully functional application set.
06. 42
These manipulation tools are;
SoTabBoxManip:
This set of manipulation tools allows the designer
to non-uniformly scale primitives. The movements
are constrained to an orthogonal grid. There is
no rotation afforded. The resultant forms were
there for classically architectural in nature.
T
The Manipulator is transformed by dragging over
the green highlighted edge conditions. Dragging
on the mid edge segment allowed for the primitive to be scaled in one direction, this is analogous to moving a face in a perpendicular plane.
Dragging on a corner segment, creates a free
face transform of the three adjacent sides.
FIGURE
06.02
Description:
SoTabBoxManip
Moving the entire object required dragging the
object but not on a vertices. This sliding movement is constrained to the axis that the picked
face defines.
06. 43
SoTransformBoxManip:
This set of manipulation tools ailow for the design-
er to uniformly scale primitives (like cubes) and
unrestrained rotation in free space. The resultant
forms were more organic/plastic in nature as a
result, as we shall see later.
The scaling function is activated by dragging the
corner boxes. The Primitive is then scaled uniformly.
Moving the object in space is much like the
SoTabBoxManip. One must click and
drag on a
face of the object and it will slide along that
plane.
FIGURE
06.03
plane.Description:
SoTransformBox
Manip
Rotation of the object is accomplished by dragging an edge. The Edge will become highlighted
and will rotate around the centroid of the object
as if one were pulling it from that point.
06. 44
SoJackManip:
This manipulation tool set allowed for easy placement of objects in accordance with a reference
axis. Height translations were facilitated by a
drag handle in the center of the object to be
manipulated. However, the lateral translations
were exceedingly difficult, especially when one
could not get hold of the edge of the primitive. It
was the most encompassing but also the most
counter intuitive to use.
The primitive can be scaled uniformly by dragging on the end boxes that appear off of the
axial spines.
The lateral translations are actuated by dragging
the edge of the superimposed symbolic box. A
Transitional plane will appear along the lateral
FIGURE
06.04
Description:
J
p
movement plane, it shows where the form will
pass through.
Vertical translations are handled by a superimposed column in the center of the primitive. This
column can be dragged down or up to facilitate
the movement of forms when the edges are not
visible.
06. 45
-U
SoHandleBoxManip
This manipulation tool set a|lows for the scaling of
primitives from the centroid bidirectionaly. It is
extremely useful when a condition for center
alignment between two forms is required, i.e. the
pediment and its associated column.
Scaling is activated by grabbing the boxes at the
ends of the axis. The two coplanar faces then
move in a mirroring fashion in relation to one
another.
Movement is
facilitated by dragging the form
from a face proper.
FIGURE
06.05
Description:
SoHandleBox
Manip
06. 46
SoTrackballManip
This manipulator tool does not afford movement.
However it allows for primitives to be rotated axially and free. It also allows for the user to alter
the centroid of rotation, such that one can rotate
the form around any point.
The movement of the axis point is activated by
T
dragging on the arrows at the junction of the
rings.
The axial rotation is activated by selecting the
ring and the primitive will slide along the activated path.
FIGURE
06.06
Description:
SoTractballManip
06. 47
Other Virtual environmental variants;
SoTextu re
One version of the Application has texture
applied to the primitives. The heightened realism
seemingly removes another level of abstraction for
the designer. The designer is afforded the ability
to react to the materiality. The computational
draw backs of this version are large. The added
memory and redraw requirements slowed the
application. This redraw retardation disrupted the
perceived smoothness of the manipulation and
navigation. It was there for difficult to work with.
SoPointLight
The point light was present in all versions. The
FIGURE
06.07
Description:
Sandstone Texture
standard viewer is equipped with a headlight
function. This function insures that all the objects
are always visible by showering the scene with
light as if it were emanating from the viewpoint
itself. With the advent of the added point light
source, the headlight can be manually switched
off. This can be done to approximate a daylighting scenario. This is a testable change that would
seem to hold promise in discovering variations of
design methods in accordance to environmental
variables. Will designers change their parti with
the added/subtracted information?
06. 48
The testing methods proper:
The study of influences of differing versions of
computational media and its ensuing feedback,
will, in this thesis, be confined to observing the
participating designers as they interact with the
application in the design process. The examination of this media-manipulation-in-process should
help inform the refinement of the application and
the investigation of finding emergent typologies
buried with-in the act of designing within differing
computational environments. It is clear that there
will not be one overriding typology due to the
simple fact that having individual designers from
different backgrounds, their styles and approaches to design will inherently be different as well.
However, I propose that there are certain natural
characteristics that become apparent due to cog-
FIGURE
06.08
Description:
Plan Villa Capra
Rotonda
Vicenza, 1552
Andrea Palladio
nitive psychological concerns that all humans
have. It is these natural characteristics that led
architects like Palladio to develop rules for designing spaces. They may not have known exactly
what it was that they were after but they reasoned that certain types or forms of spaces exhibited the qualities that they attributed to "balanced"
architecture. It is the use of manipulatable virtual
environments that could hold the key to some of
these latent characteristics.
In the course of the thesis, a set of designers will
be shown the application along with its tools and
06. 49
their operation. After being allowed an adequate time to get to know the program and its
ideosynracies, the test designers will be asked to
use the Application.
The design problem to be solved will be left up
to the individual designers, as this will clearly
demonstrate the inherent aptitudes or strengths of
each version of the application and how it would
be naturally utilized.
The architects will be tested using the ortho-constrained version and then again with the nonortho constrained version. The result of the separate tests will be saved and evaluated later.
One set of testers will be given the virtual drafting
board version. The familiarity inherent in this type
of media/interface would appear to give its
usage an edge.
Another set of designers will be given the large
scale projection/semi-immersion environment
interface.
Another set of designers will be given the Virtual
Environment in which to design.
Lastly, one small group (perhaps
just one individ-
ual) will be asked to use all the versions in an
ascending order of reality, descending order of
06. 50
abstractions. They will begin with the drafting
board and end with the VR.
The final step of the process of all of the groups
is to test the designers on what they believe that
they have created. Each architect will be asked
to draw or draft the environment that was created, WITHOUT looking at the created file for reference. This will bring to surface the intent of the
designers. Any discrepancies will there for point
to a discrepancy in understanding of form (if it is
very complex for instance), an underlying/subconsiencious emphatic relationship to the created
space or perhaps a faulty visual feedback mechanism. All of which are important factors in fully
understanding what it means to create in virtual
environments.
The files that are created will then be investigated. emergent typologies will be sought. Some of
the questions that will be investigated are: Does
one version of the application allow for true
space making? Does one version foster object
making? Isthere a particular level of complexity
that is discovered? Are there inherent restraining
angles and lines of force that can be abstracted
and utilized in differing media?
06. 51
Physical Manipulation/Interaction
Architects are familiar with the extremes of
abstraction/perception, they understand
large/full scale interactions. This familiarity comes
from experiences with built forms and their cultivated awareness of true space. Most computational interfaces are limited to a 15-17" monitor.
This simple limitation does not allow for the architect/designer to adequately examine the finer
expressions of scale and size.
In this exercise, the problematic difficulties relating
to scale will addressed in three stages. One
FIGURE
07.0 I
Description:
Architectural
Computational
Drafting board,
w/
PHEMENOarch
being the introduction of the subject/designer
into a semi-immersive environment. The second
being an illusory displacement created by an
approximation of Brunelleschi's peepshow, this
can be accomplished in two manners, one the
recreation of the peepshow, the other a large
scale simulation of the projected environment. The
third being the introduction of a fully immersive
Virtual Reality environment.
I discovered very early in the development of the
application, that with out a discourse on the physical interactiveity between the designer and the
virtual model, the infopro relationships would be
07. 52
mu
lost. Basically invalidating the experiment. So
after careful consideration, stages of designer
interaction needed to be addressed. The levels of
manipulation became 4 fold. The first starting
from the perspective of the present day architect
that is familiar with 2-d representations presented
in a drafting board environment. The next being
the introduction of a "cave" environment, that
allowed for large scale projections of the virtual
environment. The third stage is the introduction of
perspective point matching illusory space. The
final level being the introduction of a Virtual
FIGURE
07.02
Description:
PHEMENOarch
using traditional
Architectural
Media
Reality environment to the designer.
07. 53
El
The Computational Drafting Board
Architects are familiar with the abstractions that
are inherent with direct media manipulation. The
most direct example of this is the simple act of
putting pen to paper to sketch or draft. In some
ways the phenomenology of tactile feed back has
been a part of architecture since the beginning.
The German term fingerspitzgefuel, or "finger tip
feeling" adequately describes the relationship that
designers foster with their media in order to communicate their ideas and desires.
The feedback that the fingers return during the act
of drawing has been mostly lost in the computer
era. The principal reason for this disparaging discrepancy is that with most computer systems, the
FIGURE
07.03
Description:
Reflection
system
display is removed. The act of having the display
remote creates a disjunction with the designers
ability to manipulate the media.
With some simple display projectors and a calibrated image reflector, one can project a screen
upon a surface. If this imaged surface is calibrated with a stylus pen (a WACOM tablet in this
instance), one can return some of the familiar tactility of direct media manipulation to the user.
One of the side effects that I discovered with my
particular setup was that due to the nature of projection, one had to calibrated the cursor tip
FIGURE
07.04
Description:
Cursor
Displacement, 1/4"
approximately 1/4" above the actual tip of the
07. 54
El
stylus pen. This was done to compensate for the
occlusion/shadowing created by the hand and
pen. The best solution would have been a back
projection but due to the nature of the input
devices involved in this case this is impossible.
This small shift in proposed direct media manipulation, created difficulty for some of the test subjects at first. Test designers were eager to engage
with this medium immediately, sometimes even sitting down with it and experimenting without
being asked. The interface was very similar to
what they already were intuitively comfortable
FIGURE
07.05
Description:
Viewpoint of
Drafting board user
with, except for the one minor exception (the
1/4 shift) and as a result it took some real paradigm re-evaluation. After 5 minutes all the subjects were comfortable with the interface.
FIGURE
07.06
Description:
Drafting Board
Workstation and
Environment
07. 55
The Semi-Immersion
The application of semi immersive environments is
one that most computational and CAD designers
are very familiar with. The projection of perspective images upon a 2-Dimensional screen is far
from novel. The difficulty lies in the fact that most
2-Dimensional display media have no intention of
providing any kind of immersive feedback mechanisms, i.e. they are predicated upon just displaying very basic images and relationships of the virtual space. For most applications, be that engineering, industrial design, urban planning, this is
more than adequate, but for creating built inhabitable environments, like those that architecture pro-
FIGURE
07.07
Description:
S.Y Edgerton's
depiction of
Brunelleschi's first
experiment
poses, this is simply not acceptable. In order to
investigate this discrepancy I have broken this
particular problem in to two separate instru-
41
ments.
The Brunelleschi illusion
Fillippo di Ser Brunellesco (Brunelleschi, 13771446) embarked on an experiment that "marked
FIGURE
07.08
Description:
Plan of Piazza del
Duomo, what the
Panel showed
an event which ultimately was to change the
modes, if not the course of Western history"
05. 56
(Edgerton, 1975 p.3) Brunelleschi had painted a
panel that was the first painting to embody the
use of linear perspective. In order to illustrate this
point, Brunelleschi had drilled a hole in the painting, such that the viewer's eye would be constrained to the center of projection. The painting
was placed such that it faced the building that it
represented, and as the observer looked through
the back of the painting, they could see the actual scene. Then a mirror was placed in front of the
painting, reflecting the image back through the
hole to the viewer. It seemed that the real thing
was being seen.
FIGURE
07.09
Description:
Reproduction of
Brunelleschi's
As it turns out, Brunelleschi's method was very
panel, Santa Maria
effective for creating the illusion of depth. This
del Fiore in Piazza
illusion was very compelling for two reasons:
del Duomo
First, it forced the viewer to see from the very
point of projection, making the picture a projective surrogate for the scene and at the same time,
it reduced the eyes information as to the limited
depth of the image plane. There is at least one
more method of enhancing monocularly viewed
images. If a lens is inserted between the subjects
eye and the image, and if the lens approximates
the focal length of the lens that the image is taken
with (if it is a photograph that is), then the "plastic" depth that can be obtained is quite similar to
FIGURE
07. I 0
Description:
Viewpoint of
eye piece operator
w/
PHEMENOarch
binocular visions depth of field. These techniques
07. 57
can be used with computational imagery as well.
The image could be that of the virtual space and
the focal length can be approximated in accordance with FOV (field of vision).
It has been demonstrated that a proper central
projection can be mistaken for a real environment. Smith and Smith (1961) asked subjects to
throw a ball at a target in a room that they could
only see through a peephole. Half the study was
shown the actual room in the view piece, the
other half was shown a photograph taken at
exactly the same position. There seemed to be
more variation in the throws done by the subjects
that were looking at the photograph but this can
be accounted for by the direct view affording
some monocular parallax. The most interesting
aspect of the study was that the subjects had no
awareness that they were seeing a photograph
and not the actual room.
This pushed the thesis to explore the plastic effect
FIGURE
07. I I
Description:
Computer
generated rapid
prototype optics
holder
note:
This form was
generated
with the essential
optical properties
accounted for
that was created using this methodology. A relatively simple buffer was set up using an optical
convex lens. The buffer constrained the possible
view points to
just one. This view/eye point
being at the center of the computationally projected perspective. While the imagery seemed to
display the "plastic" depth that aided in the per-
07. 58
El
ception of sizes and scales, some of the test subjects complained that they were getting
headaches from viewing the monitor in such a
constrained position, and with the varying intensity caused by the interaction between the convex
lens and the CRT monitor display.
FIGURE
07. I 2
Description:
Placement and use
of eyepiece
FIGURE
07. 3
Description:
The Eyepiece
07. 59
El
The "cave"
The second stage of semi-immersion is accomplished by placing the designer at the perspective
focus point of the computer model. It should be
noted that this is just a slight altering of the
Brunelleschi method, instead of placing just the
eye of the observer at the nexus of the image,
you place the entire designer in the nexus of the
image, there by surrounding him/her.
FIGURE
The observers field of vision will almost be totally
dominated through the use of a very large prolection of the virtual environment. This type of envi-
07. I 4
Description:
Subject "M"
Designing in a
"cave" environ-
ronment has often been called a "cave". The concept of the cave is very simple, and relatively
convincing. The typical cave uses three wall projection systems, with each being directed at a different face (usually one roof and two side walls).
In the environment that I have set up, the projection will be directed towards a primary
wall/screen, and the spillover/periphery will be
projected on the ceiling, two adjacent walls and
the floor.
The subject was to use the stylus as seen in the
afore mentioned version to navigate and interact
with the virtual space, however due to technical
difficulties, the stylus was reduced to a typical
mouse and keyboard. There were many technical
difficulties to overcome with projecting from an
07. 60
SGI system. The most notable was the fact that as
true SGI projector cost in the vicinity of
+$15,000, outside of my budget. The remedy
was to use an NEC MultiSync MT800 at a lower
"video" (640x480) resolution and kick the image
through a Sirus Video board system attached to
an ONYX. This is of course analogous to swatting a fly with a sludge hammer.
Some comments from test designers have been
that the illusion is very convincing as long as one
does not move ones head around. The subtle
shifts of observation points are enough to make
the subject aware of the fact that they are still
working with a monitor and not a truly immersive
virtual space. The stylus was also a cause for
immersion breaking. The act of looking down
and/or manipulating the media of the screen pro-
FIGURE
07. I 5
Description:
Example of Spatial
Environment
created in "CAVE"
Note: the scale
references, and
attention to height
jection through a removed connection also
seemed to create a disharmony with the phenomenological tactility that the space was projecting.
An interesting side effect of this methodology was
quite unexpected. It seemed that the test designers were hesitant to engage in the media as they
had with the previous physical medias. I attribute
this to a little bit of stage fright. The simple act of
placing your exploratory design soul on screen
10-20 feet tall is a major inhibiting factor. I suppose the thought of having someone looking over
their shoulder in the dark projection hall as they
07. 61
explored their design rules and strategies was to
great for some. So as a result, in most cases the
designers produced far less "complete" work
given the same amount of time.
07. 62
The DIGIPANOMETER
As a related aside note to this thesis, another
attempt at creating a physical manipulator for virtual environments was also tried and was quite
successful.
While deliberating upon display methodologies
I
for a research grant, The unbuilt, for MOCA,
Daniel Brick , Dr. Takehiko Nagakura and myself,
/
were discussing the infusion of a new technology
called QuickTime VR with the phenomenological
characteristics of how does one experience
space. We had become very excited about the
idea that a person should interact with virtual
'11K
--------
space not only with his eyes but also with his
entire body. Before I continue, it may be helpful
for me to explain what QuickTime VR is.
QuickTime VR (QTVR) is a tool for creating and
viewing photo-realistic environments (panoramas)
and real-world objects. Users interact with
QuickTime VR content with a complete 360
FIGURE
07. I 6
Description:
Preliminary
Sketches of
Digipanometer
degree perspective and control their viewpoint
through the mouse, keyboard, trackpad, or trackball. Panoramas and objects to be viewed are
'stitched' together from digitized photographs or
3D renderings to create a realistic visual perspective.
One of the results of our conversation was the
inception of a device that would project the
07. 63
QTVR on a movable screen. This screen would
move in accordance with the movement of the
viewer. As the viewer slid to the right, the screen
would slide to the left, there by always being diametrically opposed from the viewer and allowing
the viewer to peer into the QTVR space as if it
were real space. The advantage of this methodology is obvious, by engaging the viewer in actual
body movement in order to see more of the
space, the viewer begins to gain a sense of the
dynamics of the actual space.
Dr. Takehiko Nagakura set up a team to construct
a prototype of the machine. Due to some unfortunate time constraints, I was unable to continue
with this project, even though it was shaping up
FIGURE
07. I 7
Description:
Final Sketch of
Digipanometer
to be my thesis at the time. The project ultimately
was realized by Dr. Takehiko Nagakura with
Haldane Liew and Hyun-Joon Yoo. They deserve
the credit and honor for achieving such an amazing technical feat and for their technical mastery
and craftsmanship.
07. 64
Virtual REALatiy
The third stage is the introduction of a fully immersive environment. PHEMENOarch was written primarily for use with HMD (Head Mounted Display)
and Virtual Reality.
The key commands are easily transferred to voice
commands that can be recognized with even
primitive voice recognition software (often times
built into the VR system). This simple shift in paradigms allows for the designer to be unencumbered with the minimal menu system of
PHEMENOarch.
FIGURE
07. I 8
Description:
The Forte
HMD virtual
Reality head set
The Navigation and view perspectives are routed
though the electromagnetic or sonic resonance of
the HMD. The VR helmet allows for tracking of
subtle head movements and hand's off navigation
of the virtual site. It also allows for bioptic projections that give the designer the impression of 3
dimensions.
The selections of primitives can now be relayed
in true space with direct spatial input from the
designer. 6DOF mice and /or cyber-gloves provide 3 dimensional coordinates to
PHEMENOarch. When 3 dimensional freedom is
FIGURE
07. I 9
Description:
6 DOF mouse,
accepts
3 dimensional
spatial input
compared with the other physical methodologies
used in this chapter (all of which were constrained to 2 dimensional planes), the phenome-
07. 65
nological feedback is phenomenal. One is
adding a whole other dimension (if you will pardon the pun). At this point the designer is allowed
to design with his/her whole body, a full figural
gesture. One does not need to rely upon Le
Corbusier's modular man, nor the golden rectangle or any of the other ingenious architectural
rules that experienced architects have come up
with over the centuries (that is unless you want to),
because from this vantage point you are able to
"feel" the spatial implications of composition and
form. This is the strength of PHEMENOarch.
FIGURE
The application could be linked with a haptic
device, one would then be able to add other
constraints to the modeler. Fore instance, if one
07.20
Description:
Individual using
"spacepad"' an
magnetic resonance
VR tracking system
were pushing a cube through another cube, one
would feel the resistance of the original cube
resisting this intrusion. The ramifications of this simple act is that the designer could begin to understand the whole object and its interaction with
other existing objects without having to be at a
vantage point to "see" the interactions. Of course
there could be a myriad of other links that could
be manifest in such a device, like feeling the
forces which a certain structural configuration
could resist, the simulation of air flow patterns
through a space which could be felt, the actual
molding of forms as though they were clay, reinventing the "fingerspitzgefuel" tactility of true physical building.
07. 66
U'-
All these factors aid in the intent of creating an
application (PHEMENOarch) which would stimulate right brain infopro. The designer is allowed
to concentrate upon the task of creation and is
not pulled out of the immersion by having to
switch cognitive paradigms to perform different
tasks.
PHEMENOarch is ideally suited for the VR environment for the previous reasons and because it
is written in the native language of Open
|nventor, of which VRML (virtual reality modeling
language) is based. However due to management and proprietary restraints, I was unable to
FIGURE
07.2 I
Description:
This particular VR
headset also incorporates full stereo
surround sound
fully test out the application with this equipment.
My preliminary research was quite encouraging
and is an avenue that I believe should be
explored more in-depth.
07. 67
The Test Experiments
The thesis became divided half way through as to
which research methodology should be
employed. There are several different research
strategies that could have been employed.
Experiment, survey, archival analysis, history and
Case study are the most commonly used research
methodologies. In the beginning, the experiment
methodology seemed to produce the greatest
results. However as the testing progressed, it
became apparent that the research methodology
which this thesis should employ was one of Case
Study.
reflections on Abstractions
While the experimental evidence would require
more empirical testing to validate my findings,
there are some interesting preliminary implications.
08. 68
Methodology;
Experiment (the Blue phase)
Early in the testing phase of this thesis, the tests
were geared for an experiment research strategy.
(b)
F
The test subjects were to be given limited versions
of PHEMENOarch. The "blue" model is a perfect
example of this stratagem. In the preliminary versions of PHEMENOarch, the primitive forms were
all a shade of blue due to the added depth that
(C) q
3
they seemed to convey on the SGI monitors.
In this experiment, the intent was to discover
whether or not an orthogonal constrain system
would facilitate orthogonal spaces. The hypothesis was that the spaces would have orthogonal
elements, by the very nature of the constrained
program but the spaces that were created could
be abstracted to reveal a "plastic" spatial form.
This form could then be reinterpreted and demonstrated as a series of abstractions that could be
in-turn brought back to traditional methods of
FIGURE
08.0 I
Description:
The Golden Section
architectural visualization. I.E. one could interpret
a new set of basic cognitive rule systems and
forms like the "golden section".
The test subject was asked to design two different
spaces. One was to be a public space the other
was to be private in nature.
Out of the four test subjects, three of them began
with the private space first. The construction of the
08. 69
floor plane and ceiling plane seemed to be paramount to their subsequent relationship building.
The walls were manipulated with an emphasis on
the geometry at the apex or junction points of
horizontal and vertical members. If a window
were to be placed in the wall, there was great
attention placed upon the articulation of the surrounding pieces. In a few cases, the concept of
built ruins (AKA. Louis Kahn) seemed to emerge.
The significant articulation of depth of the width
of the "private" space walls seemed to be a derivative of the belief that this environment should
seem, as one tester put it, "safe and protected".
This may seem like an intuitive notion of the phenomenology of space, but it is exciting to see it
articulated in such a precise manner with in the
FIGURE
08.02
Description:
Window
penetration in
private space
note the depth of
the wall section
test environment.
The careful elaboration of the transitional gate or
passage way between the public and private
space seemed to solidify architectural tactility. In
almost all the cases (3 of 4) the passage
between two spaces were crafted such that one
could not see from the public space into the private. This was usually accomplished by making
the passageway a corner of the public space,
and then by making the passageway relatively
deep. This made it difficult for an individual that
FIGURE
08.03
Description:
entrance to private
space from public
area
is residing with in the public space to gain a
08. 70
view access point into the private space unless
he/she were entering the room. Inone instance,
a designer/tester created a "window" from the
private space to the public space that, due to its
elevation, allowed for a voyeuristic one way view
over the public area.
Another interesting phenomenon which occurred
was completely unexpected. The interior of the
private space was not rectangular. The wall of
the interior private space was angled back from
the wall at (depending on the subject) 12 -18
degrees. This effect may be the result of some
error in the manner that the computer displays the
virtual environment back to the designer. The
other possibility is that there is some cognitive
08.04
Description:
stepping back of
interior of private
space wall.
FIGURE
need to want to tweak the perspective of the private space such that it displays some as yet
undiscovered characteristic. Both of these hypothesis will need further investigation in order to
unveil the true reason behind this interesting construction.
note the implied
angel in relation to
the orthogonal
"step"
In listening to the designers discuss what they
were doing and why, it was noteworthy that their
conversations seemed to revolve about the subtle
play of view obscuring and demonstrating, and
what the space should "feel" like.
08. 71
-I-----
MUMAN
The public spaces provided some what less definable results.
The definition of "public space" seems to be to
encompassing for there to be enough coordination between the relatively small sampling that I
was able to achieve. There were some startling
results though.
The relationship between all the public and private spaces seemed to be one of a constant
ratio. The private space, regardless of its placement in the over all context, was always between
1/3rd and 1/2 of the size of the public space.
FIGURE
08.05
Description:
Public space
example
This seems to validate the architectural "golden
section". This ratio is an even more amazing
since the "true" scale of one subject's project varied greatly with the scale of the others. When a
designer first enters the virtual space, the tester is
confronted with a singular cube. Some testers
took this cube as a unit scale and they would
"back away" from the cube until it was of an
"appropriate scale" to be used. Other testers,
FIGURE
immediately scaled the block to an "appropriate
scale" from their current "home" viewpoint. This
second group of testers created very "small" mod-
08.06
Description:
Public / Private
space
Ratio
els in relationship to the first group, because they
were constantly scaling the primitives down to
their perceived level. One tester, even complained that she wished that "the cubes would
come in at a smaller scale"
08. 72
There were some expected yields from the public
space experiments.
In some of my earlier works and investigations
A
(before thesis), some of the preliminary discussions of designing within a perspectival space
B
C
led me to investigate certain aspects of spatial
D
relations. Many of the same concerns manifested
themselves here. For example, if one is designing
a colonnade in a singular perspectival space.
The typical perceptual mistake is to believe that
the columns are equally spaced and are of equivalent sizes, when in reality the sizes are usually
increasing (dramatically) and the spacing is also
increasing (dramatically) as the columns move far-
aA
bB
c C
d
eD
Percieved
Columns
ActualColumns
ther away from the creation camera/designer
viewpoint. This is an inherent danger with a one
point perspective, as it is very difficult to gauge
depth accurately with this methodology. This perceptual mistake is also possible in
PHEMENOarch, if the designer does not move
FIGURE
08.07
Description:
Column diagram
Vantage Point
Design dangers/
opportunities
his/her vantage point in order to reevaluate the
space. In some instances, this perceptual vagueness can be utilized by the designer too great
effect. One designer created a rhythm by making
the blocks sit on each other all the same size, but
as he stacked the blocks closer to the ceiling of
his enclosure, he specifically mad the blocks
smaller and thinner such that the observer from
the ground would feel that the space was far
greater than it actually was. When asked about
08. 73
this later, he replied that he wanted to see if "anyone noticed". Another designer wanted to create
a set of stairs, as he finished creating them from
one vantage point, he moved to another corner
of his construction. He was amazed to note that
the stairs were "not even close" to where he had
believed that they were. He became excited by
the development and began to experiment with
crating forms that appeared to be one thing from
one side and then changed dramatically when
viewed from another (unfortunately this model
was lost due to an unexpected core dump).
08. 74
Methodology: Case Study
1,as researcher, was looking for certain characteristics which I had originally postulated, a
hypothesis mentality. The difficulty in bringing
these hypotheses to fruition became apparent half
way through the testing phase. The preconceived
notions that I as researcher, designer and developer had come up with seemed to impede the
exploration of what was this program and physical interface affording the designer. The decision
of following an experimental methodology
seemed to produce decent results, however I was
not convinced that they were demonstrating a
"true" usage of the provided medias. It was also
evident that the inherent control over behavioral
events that was required with the experimental
method, was in a way impeding the larger pic-
FIGURE
08.08
Description:
Advanced Case
Study
Subject "B"
using
PHEMENOarch
on INDY SGI
ture of how, why and what do designers create
in Virtual free space.
The range of testers was broadened to include
several varying backgrounds. PHEMENOarch
was tested with different end-users, they ranged
from Investment Bankers, to Engineers, from
Students of Architecture, to accredited Architects,
from programmers, to artists.
It was at this point that the test/designers were
given a 10 minute demonstration of the capabilities of the program and then they were told to
08. 75
create anything that they wished. The intent was
to see if the "immersive qualities" of the software
would entice the testers into using it as an architectural space former, or would they utilize it as
an object creator. The results were not as expected. A major discrepancy emerged that is difficult
to understand.
The testers that had little or no CAD (computer
aided design) experience, seemed to pick up the
program basics very easily, they had a little difficulty adjusting to the differing mindset that was
required, but after 15-20 minutes of working with
PHEMENOarch, they were able to create some
very architectural readable pieces and forms.
Many of these subjects did not even have an
architectural background and yet even with a
very limited design vocabulary, they were navigating through their sites to "envision"
how one
would approach their "buildings". This was the
key, in some manner, they were navigating
through the built forms to evaluate them rather
than navigating around the built forms for evalua-
FIGURE
08.09
Description:
Subject "A"
epiCe
expenience
tion.
Engineer
... this is how you approach the entrance.... like
that,... / wanted this to be the grand "facade",...
god help me, I feel like an archy..
subject 'A", engineer
08. 76
We live in a 3 dimensional world, and yet we
have great difficulties thinking in true 3 dimensions. This application forces you to think in 3
dimensions. ... It is definitely a mind shift, if you
will.... It takes some getting use too, but after 5
min. I was really having fun with it! You should
really think about using this to teach architects
about space and relationships..... It really
opened up my eyes as to how to visualize distances, I never thought about it that way.....
subject "T", on investment banker
The test designers that had "some to moderate"
experience with CAD systems and design,
seemed to have fundamental difficulties with
manipulating the built space. In most cases, this
group consisted of experienced architects that
were used too envisioning spatial relationships.
FIGURE
08. I 0
Description:
subject "T"
No CAD
experience
investment banker
08. 77
qi-
Here is one sample case studies notes;
Subject "", architect and architectural graduate
student
e Subjects first inclination isto create "planes"
these thin forms are being created at a distance
such that fine control can not be achieved, the
program crashes several times as forms are force
to infinity (infinitely large, or infinitely small, ll
have to fix that bug)
e The manner of spatial manipulation is not one
of moving though the site but rather by maintaining one perspective and "finishing the form",
then moves to see if the forms are correct.
* "'s" movement characteristics are that of more
vertical moves than transversal movement, approximately 4 to 1
FIGURE
08. I I
Description:
subject "J"
Moderate CAD
experience
Architect
* " uses the auxiliary horizontal and tilt movement manipulators around the edge of
PHEMENOarch rather than the main ones provided by the implementation of the SoWalkViewer.
aside note: "" wanted to "walk" through the
space after it was completed.
Note: the final form
is of spatial nature,
however, the
methodology was
one of plan
abstraction
e Requested plan and section views as, subject
"was used to it" as an architect.
* Wanted to "ump" from viewpoint to viewpoint
(much like looking at different elevations)
o Subject elevates the viewpoint position, and tilts
down, there by recreating a plan view.
* Crashes software again (more infinity difficul-
08. 78
-1
ties), / recovered the work.
When asked about problems and comments:
subject believed that the navigation tools needed
work.
"architectsare used to thinking in space and
drawing in plan, with this... you have to think in
plan and draw in space."
subject felt that it would be "easier" if immersed in
VR, "would help a lot"
This is an interesting phenomenon that I would
have liked to purse further given more time.
The "experienced" CAD group had very little difficulties with adapting to the new environment
offered by PHEMENOarch. It was a shift in view-
Description:
Subject "M"
point paradigms, and still required a few minutes
of acclimation. Once this stage was passed, they
High CAD
Experience
were able to use the application to its fullest
extent. The were the true space builders. 3 of the
test designers requested copies of the application
FIGURE
08. I 2
Architect
to continue with their spatial explorations.
08. 79
El-
Some of their comments;
Boy isthis easier than AutoCAD..... I like it!
-subject "K", practicing architect, avid
AutoCAD user
... the subtitle in the manner with which one
manipulates the objects is very much the quintessential heart of the matter. There are no dialogue
boxes,... nothing pops out at you.... it is tactile
and it beckons exploration.... I believe that it isa
piece isabout experience and not computation.
-subject "S", architect and Architecture
graduate student/teacher
FIGURE
08. I 3
Description:
Subject "K"
AutoCAD
Architect
Many of the same discussions and abstractions
appeared with this research methodology and it's
accompanying group of designers, as did with
those from the earlier "experimental Blue group".
The most relevant are as follows:
FIGURE
08. I 4
Description:
Subject "S"
CAD experienced
Architect/Grad.
08. 80
Edge Boundaries
In the inception of boundary conditions, i.e.
walls and enclosures, the tendency or inclination
is to break up edge conditions. This may be due
to some of our architectural training, the interest
of exploring the dislocating of boundaries. Given
this reasoning it is not unexpected for this condition to occur but what is unexpected is the regularity with which this condition occurs. The basic
wall geometries can be decomposed into a sim-
FIGURE
08. I 5
Description:
Sectional Tracing
of Edge Boundaries
ple boundary spine, this spine becomes exaggerated when a perpendicular wall intersect the original wall.
FIGURE
08. I 6
Description:
3-Dimensional
representation of
abstractions
08. 81
El
The spine is apparently based upon eye
height/view point. This can be approximated by
creating a cylinder that runs along the boundary,
with the centroid of the cylinder at eye height.
view Height
Hi
Models superimposed
view Height
Hi
Raw Extrapolated Data Curve
A
A
2A
view Height
Hi
-
2A
2A
Abstraction / Grammar
FIGURE
08. I 7
Description:
Sectional Diagrams
of spatial grammar
08. 82
mm
Light
The addition of other factors has certain effects
on the creation of the space. When a uni-directional light source was added to the design environment, the designers had a tendency to alter
their design moves. If they were facing a wall
that was washed with light, their reactions were
for the most part, to create subtle juxtapositions of
forms in order to create the sought dynamic. This
was drastically different from the opposing situation. The same designers, when they were facing
FIGURE
08. I 8
Description:
Example of the
subtle facade
manipulations in
the "sunlit"areas
forms that were backlit (not in direct light), the formal moves became more aggressive. The relative
juxtapositions were on a larger scale. This can be
explained I believe by the phenomenological
reaction to the limited light levels, in order for the
designers internal infopro model to be realized
the forms had to take on different proportions and
scale to develop the same visceral reaction as the
omni-lit model environment. In this application, the
modeling of light was very basic, but I believe
that it is obvious that the necessity of modeling
light in a more accurate fashion has been demonFIGURE
strated and should be investigated further.
08. I 9
Description:
Exaggerated
architectural moves
in the "sun
occluded" areas
08. 83
Transitional Boundaries
The boundary between formal spaces were exaggerated by almost every designer. The entrance
usually became an event unto itself. The transitional spaces were often created with a depth relationship that had a proportional systems.
A specific condition that produces measurable
results is an entrance on the corner of an
enclosed space. The rectangularity of the space
is often forced into a trapezoidal form. The reoccurring angle of this final side seems to be
FIGURE 08.20
Description:
between 12 and 18 degrees. This time the
designers were asked to draw their intentions,
Boundary between
spaces
Transitional
and a quarter of them did not realize that they
had done it.
Private Space
Public Space
FIGURE
08.2 1
Description:
Entrance Diagram
08. 84
Trouble with Textures
When the test designers were on a machine that
had enough power to deal with the texture
redraws, the amount of primitives and/or the
completeness of the designs actually was
decreased compared with the designers that
were not given textures. The designers seemed to
"work" the visual representations of materiality to
such an extent that it seemed to the observer that
they had lost the direction of the overall space.
FIGURE
08.22
Description:
Abstractions
maintained
with out textures
This seems to leads to a verification of the concept that design must begin in stages of abstraction, with verisimilitude being the last abstractive
barrier/stage. It may also be a question of the
test designers having never been exposed to this
type of virtual media creation and that they simply want to experiment with it. More testing has
to be done to see conclusively whether this tendency will continue with more media
mature/savvy subjects.
FIGURE
08.23
Description:
Difficulties with
implied realism
with textures
08. 85
S
m
Conclusions:
I believe that this thesis is a small tentative step
into a new direction for architectural design
methodology. This computational medium has
potential and must be explored further. Its impact
may be as great as the impact that perspectival
drawing had on Renaissance architecture.
/
/
The application has demonstrated great promise
as a tool for space manipulation and space
building. I would have liked to see more experimentation done in the field of Virtual Reality, to
truly push the phenomenological concerns to their
ultimate limits. There are a few small functions
that I would have liked to add to the program,
the most notable of which is Booleans. With a
complete tool set, one could begin to push all the
avenues of architectural spatial thought. One
could think in terms of subtractive formal gestures
as well as the additive ones that are currently provided. In order to make PHEMENOarch a true
design tool for architectural creation and not just
an experimental tool of cognitive spatial exploration, a number functions must be added.
Probably the most notable of which are copy and
paste tools, and more importantly an "undo". The
addition of traditional abstractions like plans and
sections would also have to be carefully integrated. The color palate in PHEMENOarch was for
me the most problematic and least successful of
the tools. The ultimate intent was to have the
architect never be presented a "menu". Though I
ran out of time, the color palette manipulator was
to be a push and pull manipulator that would
A-B
'-'-7
A
B-A
4<7
-. '~'-
-~
/
/
V
FIGURE
09.0 I
Description:
Examples of
Subtractive
Boolean functions
09. 86
slide you through the color chart, in 3 dimensions. I know that this can be accomplished with
a simple 3vect.
The manipulation and understanding of scale is
another issue that needs to be addressed. I was
working on a visual displacement ruler that would
demonstrate scale distances between two chosen
objects. The necessity for spatial scale was
demonstrated by the designers that made their
models at a super small scale. Perhaps the addition of a texture mapped "Modular Man" would
be all that was needed.
The method of immersive form manipulation has
proven to be a valid one. The relative ease that
individuals with little to no architectural or artistic
background were able to grasp and embrace the
complex concepts of 3 dimensional modeling is
astounding. A result that was totally unexpected.
There is still some concern and confusion over
why trained architects and average CAD modelers had such difficulty with the concept of the software. It is an avenue that will require more
thought. It seems strange that spatial literate
designers can not change their mind set to visualize space in a different manner.
tic
FIGURE
09.02
Description:
Le Corbusier
Modular Man
The area of architectural design and computation
is just developing. The research into shape grammars and parametric design is very exciting and
it holds some very economically lucrative and theoretically challenging underpinnings. The research
into these areas are necessary. I would believe
that the next level of this program would have
parametric constraints built into it. For example, if
one moved the walls apart further than would be
structurally feasible, the program would not allow
the move to continue. The architect could then
specify a stiffener or some other structural rein-
09. 87
forcement and be able to continue. For a "real
world" architect, it would also be interesting if the
program returned a cost analysis while the
changes were being made, such that the architect could be able to balance the design to the
given budget. With the option of turning it off.
Another avenue for exploration with these tools, is
that of architect training. This tool as it stands
now, affords a mentor or professor a manner with
which to demonstrate spatial ideas and forms, as
well as forcing the student user to visualize forms
in 3 dimensional space. It would also be exciting
to study how children might react to this empowering environment building.
The architect is the creator of ideas, the challenger of normative mass mentalities and thinking.
The architect is the mirror of society. The search
for a "perfect" architecture is fruitless, because
architecture is about humanity and must there for
change and grow with us. It must not be complacent. We must continue to develop tools that
afford the architect a new vista to see and
explore.
09. 88
;; This is the PHEMENOarch program.
;; It is intended to be used to build architectural spaces
;; FROM the INSIDE, i.e. as an inhabitant of the space.
;; 00000000000000000
;; 00000000000000000
;; 00000000000000000
;; this defines the viewer
;; 00000000000000000
;; 00000000000000000
;; 00000000000000000
(define viewer (new-SoXtWa kViewer))
(-> viewer 'show)
(-> viewer 'setViewing 0)
(define scene-root (new-SoSeparator))
(-> scene-root 'ref)
(define material-editor (new-SoXTMaterialEditor))
-> material-editor 'setTitle "Material Editor")
-> material-editor 'show)
;
;
00000000000000000
00000000000000000
;; 00000000000000000
; the sun light is approximated with this function
;; 00000000000000000
;; 00000000000000000
ci. 89
;; 00000000000000000
(define side-light-light (new-sopointlight))
(-> scene-root 'addchild side-light-light)
(-> (-> side-light-light 'on) 'setvalue ])
(-> (-> side-light-light 'location) 'setvalue 20 20 101)
;;
;;
;;
;;
00000000000000000
00000000000000000
00000000000000000
this code defines how objects are selected and how manips are transfered
;; 00000000000000000
;; 00000000000000000
;; 00000000000000000
(define (selection-cb user-data path)
(let ((object (-> path 'getTail))
(sep (SoSeparator-cast (-> path 'getNodeFromTail 1)))
(display "Selection Callback: Object
(display object)
")
(display " was selected.")
(newline)
(display "Selection Callback: Object Type is ")
(display (-> (-> (-> object 'getTypelD) 'getName) 'getString))
(newline)
(if (= 1 (-> object 'isOfType (SoCube::getClassTypeld)))
(let ((the-sphere (SoCube-cast object)))
(display "Selection Callback: Picked a cube.")
(newline)
;; additional operations with sphere
))
(if (= 1 (-> object 'isOfType (SoTransform::getClassTypeld)))
(let ((the-sphere (SoTransform-cast object)))
(display "Selection Callback: Picked a transform.")
c1. 90
(newline)
00000000000000000
;o00000000000000000
;o00000000000000000
;; additional operations with cube/eye
; sotabboxmanip, sotransformboxmanip
; socenterbalimanip, sohandelboxmanip
; 00000000000000000
;i00000000000000000
;;
00000000000000000
(define new-manip (new-sotabboxmanip))
(-> new-manip 'replaceNode path)
(display "Selection Ca||back: Object Type Of Sep's Child 1 is "
(display (-> (-> (-> (-> sep 'getChild 1) 'getTypelD) 'getName)
getString
(newline)
(-> material-editor 'attach (SoMaterial-cast (-> sep 'getChild 1)))
(define (deselection-cb user-data path)
(let ((object (-> path 'getTail))
(sep (SoSeparator-cast (-> path 'getNodeFromTail 1)))
(display "Deselection Callback: Object
(display object)
(display " was deselected.")
(newline)
")
(display "Deselection Callback: Object Type is ")
(display (-> (-> (-> object 'getTypelD) 'getName) 'getString))
(newline)
Ci. 91
(if (= 1 (-> object 'isOfType (SoSphere::getClassTypeld)))
(let ((the-sphere (SoSphere-cast object))
(manip (SoTransformManip-cast (-> sep 'getChild 0)))
(display "Deselect Callback: Picked a sphere.")
(newline)
(define new-trans (new-sotransform))
(-> sep 'replaceChild 0 new-trans)
(-> (->
(> (->
(-> (->
(-> (->
new-trans 'translation) 'setValue (-> manip 'translation))
new-trans 'rotation) 'setValue (-> manip 'rotation))
new-trans 'scalefactor) 'setValue (-> manip 'scalefactor))
new-trans 'scaleorientation) 'setValue (-> manip 'scaleorien-
tation))
(-> (-> new-trans 'center) 'setValue (-> manip 'center))
(if (= 1 (-> object 'isOfType (SoTransformManip::getClassTypeld)))
(let ((manip (SoTransformManip-cast object)))
(define new-trans (new-sotransform))
(-> manip 'replaceManip path new-trans)
(define (pick-filter-cb user-data picked-point)
(let ((path (-> picked-point 'getPath)))
(let ((node (-> path 'getTail)))
(if (= 0 (-> node 'isoftype (sotransform::getclasstypeid)))
(let ((new-path (-> path 'copy 0 (- (-> path 'getLength) 1)))
(sep (SoSeparator-cast (-> path 'getNodeFromTail 1))))
(display "Pick Filter: Non-Transform at the end of the path")
(newline)
C1.
92
(-> new-path 'append (-> sep 'getChild 0))
new-path
(begin
(display "Pick Filter: Transform at end of path.")
(newline)
path)
(define root (new-SoSelection))
-> scene-root 'addchild root)
> -> root 'policy) 'setValue SoSelection::SHIFT)
> root 'addSelectionCallback
(get-scheme-selection-path-cb)
(void-cast (callback-info selection-cb)))
> root 'addDeselectionCallback
(get-scheme-selection-path-cb)
(void-cast (callback-info deselection-cb)))
> root 'setPickFi IterCa Ilback
(get-scheme-selection-pick-cb)
(void-cast (callback-info pick-filter-cb)))
; 00000000000000000
;; 00000000000000000
; 00000000000000000
;; KEYBOARD EVENT STUFF
; 00000000000000000
; 00000000000000000
; 00000000000000000
(define (keypress-cb user-data event-callback)
(let ((event (send event-callback 'getEvent)))
(cond ((= 1 (SO_KEYPRESSEVENT event c))
(display user-data)
(create-new-eye root)
c i. 93
(send event-callback 'setHandled))
((= 1 (SO_KEY_PRESS_EVENT event d))
(display user-data)
(create-new-eye2 root)
(send event-callback 'setHandled))
(( 1 (SO_KEYPRESSEVENT event g))
(display user-data)
(create-new-eye3 root)
(send event-callback 'setHandled))
((= 1 (SO_KEYPRESSEVENT event s))
(display user-data)
(print-scene "newfile.iv" scene-root))
((= 1(SOKEYPRESSEVENT event b))
(display user-data)
(newline)
(display "Here's where I autocenter the centerball dragger.")
(newline)
;; 'root' is the SoSelection Node
(display "Number of selected paths is ")
(display (-> root 'getNumSelected))
(newline)
(if(< 0 (-> root 'getnumselected))
(let ((path (-> root 'getPath 0)))
(let ((object (SoTransformManip-cast (-> path 'getTail))))
(display "Autocenter.")
(newline)
(cond
((= 1 (-> object 'isOfType
(SoCenterBallManip::getClassTypeld)))
(display "CenterBallManip - Center it.")
(newline)
(display "This doesn't work. Disbaled it.")
(newline)
ci. 94
MEMEMEMEMBEEMP
I
--- --
-
;; (-> (-> object 'center) 'setValue 0 0 0)
((= 1 (SOKEYPRESSEVENT event a))
(display user-data)
(newline)
(display "Here's where I toggle the manipulator type.")
(newline)
;; 'root' is the SoSelection Node
(display "Number of selected paths is ")
(display (-> root 'getNumSelected))
(newline)
(if (< 0 (-> root 'getnumselected))
(let ((path (-> root 'getPath 0)))
(let ((object (-> path 'getTail)))
(display "DO IT!")
(newline)
(display "Switch the Manip: Object Type is ")
(display (-> (-> (-> object 'getTypelD) 'getName)
'getString))
(newline)
(cond
((= 11 -> object 'isOfType
(SoTabBoxManip: :getClassTypeld)))
(display "TabBoxManip")
(newline)
(define new-manip (new-sotransformboxmanip))
(-> new-manip 'replacenode path))
((= ] (-> object 'isOfType
(SoTransformBoxManip::getClassTypeld)))
(display "TransformBoxManip")
(newline)
(define new-manip (new-socenterballmanip))
(-> new-manip 'replacenode path))
c i. 95
1 (-> object 'IsOfType
(Socenterba lManip::getClassTypeld)))
(display "SoJackManip")
(newline)
(define new-manip (new-SojackManip))
(->
new-manip 'replacenode path)
((= 1 (-> object 'isOfType
(SojackManip::getClassTypeld) )
(display "CenterBa|lManip"
(new ne
(define new-manip (new-sohandleboxmanip))
(-> new-manip 'replacenode path))
((= 1 (-> object 'isOfType
(SoHand leBoxMa nip:: getClassTypeld)))
(display "HandleBoxManip")
(newline)
(define new-manip (new-sotabboxmanip))
(> new-manip 'replacenode path)))
(begin
(display "Nothing is selected. Do nothing.")
(newline))
(send event-callback 'setHandled))
(define ev-cb (new-SoEventCallback))
> ev-cb 'addEventCallback
(SoKeyboardEvent::getClassTypeld)
(get-scheme-event-callback-cb)
(void-cast (callback-info keypress-cb))
ci. 96
> root 'addchild ev-cb)
; 00000000000000000
; o0000000000000000
;; 00000000000000000
; EYE/Primitive Cube CREATION PROCEDURE
; 00000000000000000
; 00000000000000000
; 00000000000000000
(define (create-new-eye geometry-root)
(define eye (new-soseparator))
(-> geometry-root 'addchild eye)
(define eye-trans (new-sotransform))
(-> eye 'addchild eye-trans)
(define eye-mat (new-somaterial))
(-> eye 'addchild eye-mat)
(-> (-> eye-mat 'diffusecolor) 'setvalue 1 1 1)
(define texture (new-SoTexture2))
(-> (-> texture 'filename) 'setvalue "img.rgb")
;; filename.rgb should be replaced with the appropriate filename
;; I'm not sure but 01 may only understand SGI rgb format images.
(-> eye 'addchild texture)
(define eye-sphere (new-Socube))
(-> eye 'addchild eye-sphere)
(if (< 0 (-> root 'getnumselected))
(let ((path (-> root 'getPath 0)))
(let ((object (SoTransformManip-cast (-> path 'getTail))))
(display "Create-New-Eye: Setting the transformation.")
(newline)
cl. 97
(define vec (-> (-> object 'translation) 'getValue))
(define vec1 (new-SbVec3f 1 1 1))
(-> (-> eye-trans 'translation) 'setValue
vec 'operator+ vec 1)))))
eye
(define (create-new-eye2 geometry-root)
(define eye2 (new-soseparator))
(-> geometry-root 'addchild eye2)
(define eye-trans2 (new-sotransform))
(-> eye2 'addchild eye-trans2)
(define eye-mat2 (new-somaterial))
(-> eye2 'addchild eye-mat2)
(-> (-> eye-mat2 'diffusecolor) 'setvalue 0.5 0.5 0.5)
(define texture2 (new-SoTexture2))
(-> (-> texture2 'filename) 'setvalue "Conc.rgb")
*.rgb should be replaced with the appropriate filename
;; I'm not sure but 01 may only understand SGI rgb format images.
(-> eye2 'addchild texture2)
(define eye-sphere2 (new-Socube))
(-> eye2 'addchild eye-sphere2)
(if (< 0 (-> root 'getnumselected))
(let ((path (-> root 'getPath 0)))
(let ((object (SoTransformManip-cast (-> path 'getTail))))
(display "Create-New-Eye: Setting the transformation.")
(newline)
(define vec (-> (-> object 'translation) 'getValue))
(define vec] (new-SbVec3f 0 2.1 0))
(->
eye-trans2 'translation) 'setValue
vec 'operator+ vec1)))))
c I.
98
eye2
(define (create-new-eye3 geometry-root)
(define eye3 (new-soseparator))
(-> geometry-root 'addchild eye3)
(define eye-trans3 (new-sotransform))
(-> eye3 'addchild eye-trans3)
(define eye-mat3 (new-somaterial))
(> eye3 'addchild eye-mat3)
(->
(>
(->
(>
(->
(->
(->
(->
eye-mat3
eye-mat3
eye-mat3
eye-mat3
'diffusecolor) 'setvalue 0 0.2 0)
'transparency) 'setvalue 0.6)
'specularcolor) 'setvalue 0 1 0.2)
'shininess) 'setvalue 0. 1)
(define texture3 (new-SoTexture2))
(-> (-> texture3 'filename) 'setvalue "img.rgb")
;; filename.rgb should be replaced with the appropriate filename
;; I'm not sure but 01 may only understand SGI rgb format images.
(> eye3 'addchild texture3)
(define eye-sphere3 (new-Socube)
(-> eye3 'addchild eye-sphere3)
(if (< 0 (-> root 'getnumselected))
(let ((path (-> root 'getPath 0)))
(let ((object (SoTransformManip-cast (-> path 'getTail))))
(display "Create-New-Eye: Setting the transformation.")
(newl ine)
(define vec (-> (-> object 'translation) 'getValue))
(define vec] (new-SbVec3f 1 1 1))
(-> (-> eye-trans3 'translation) 'setValue
vec 'operator+ vec] )))))
eye3
c I. 99
(create-new-eye root)
;00000000000000000
;00000000000000000
;; PROCEDURE TO WRITE SCENE FILE
; 00000000000000000
;; 00000000000000000
(define (print-scene filename scene-root)
(define write (new-SoWriteAction))
(-> (-> write 'getOutput) 'openFile filename)
(-> (-> write 'getOutput) 'setBinary 0)
(-> write 'apply scene-root)
(-> (-> write 'getOutput) 'closeFile)
(-> viewer 'setscenegraph scene-root)
(gc)
; 00000000000000000
;; 00000000000000000
;; 00000000000000000
;; To open an .iv file named "filein.iv" un comment this!
; 00000000000000000
;; 00000000000000000
;(define root-in (read-from-inventor-file "filein.iv"))
;(-> root 'addchild root-in)
c i. 1 00
- - -
TO SET UP THE ENVIRONMENT FOR PHEMENOarch
The following set up is taken from:
http://vismod.www. media. mit.edu/courses/cgw97/
This is definatly a web page to investigate. It taught me a lot, and I can not recommend it
highly enough.
The development environment we will use on Athena is a version of Scheme (SCM)
which has an Open Inventor binding. This allows you to write Scheme programs which use
the Open Inventor 3D graphics toolkit.
Note that currently Open Inventor is only available on Athena SGIs. Therefore, you will
have to find an Indy in order to be able to work with Inventor.
To get started, add the following line to the end of your ~/.environment file:
add iap-cgw
# Current location of IvySCM
Add the following lines to the end of your -/.cshrc.mine file:
if ($hosttype
==
sgi) then
limit coredumpsize 0
# Don't let anything dump core
setenv SCHEMEINITPATH /afs/athena.mit.edu/course/other/iap
cgw/lib/scm/
setenv SCHEMELIBRARYPATH
/afs/athena.mit.edu/course/other/iap-
cgw/lib/slib/
OA. 1 0 1
setenv LDLIBRARYPATH /afs/athena.mit.edu/course/other/iap-
c
cgw/lib
endif
Note: you will need to log out and log back in for the above changes to take effect.
Although you can run the Scheme interpreter, ivyscm, directly from an Athena prompt, it
is strongly recommend that you run it from within emacs. To set up ivyscm within emacs, add
the following lines to the end of your -/.emacs file (or create one if you don't have one):
(setq load-path
(cons "/afs/athena.mit.edu/course/other/iap-cgw/elisp"
load-path))
Use scheme-mode for files with .scm suffix
(setq auto-mode-alist
(append '(('\\.scm$" . scheme-mode)
auto-mode-alist))
Autoload run-scheme from file cmuscheme.el
(setq scheme-mode-hook
'((lambda () (autoload 'run-scheme "cmuscheme"
"Run an inferior Scheme" t))))
(autoload 'run-scheme "cmuscheme"
"Run an inferior Scheme process."
t)
(setq scheme-program-name "ivyscm")
This will set up emacs with IVY and Scheme.
OA. 1 02
TO USE PHEMENOarch
I would like to add just a few step by step commands that will be helpful
for one to evaluate and run PHEMENOarch, especially if one is new or
unfamiliar with the computation environment.
Once you have the emacs window open, type with in the buffer:
M-x run-scheme
(M stands for meta, this is the "Alt" key)
While you can work by typing directly into the interpreter's buffer, it is
much more convenient to work in a separate buffer. To create two windows, one for the evaluating scheme buffer and one for the evaluated
PHEMENOarch application, just type:
<Ctrl> x 2
Move your cursor to the upper window and click once to activate the window. Then load the file PHEMENOarch.scm by typing:
C-c C-
scheme-load-file, then PHEMENOarch.scm
Or One can also go to the menu bar and "open file" like a traditional program.
Evaluate the file by going to "home" (the top) and type:
OB. 1 03
<Ctrl> <space bar>
(mark set)
Then go to the end and type:
<Ctrl> c <Ctrl> r
(evaluate region)
If you make a mistake type
<Ctr|> g
to stop any errant processes.
This should get you up and running with the application. I would suggest
taking some Emacs courses from Athena if you want to get more in depth.
You should see the Walkviewer appear with a cube in the middle.
The viewer will have certain buttons down it left side. The pointer refers to
a selection pointer, with this selected, you can move and alter objects.
useful commands with in PHEMENOarch.scm in this mode are:
a
=
to toggle methods of manipulating the primitives
to add a "sandstone" cube, please note that if a cube is selected
already it will place the new cube in relation to the selected one. If there
c
=
is no cube selected, the cube will appear at the origin.
d =to add a different cube (I know "real original"), this one will have a
concrete texture. Again if a primitive is already selected, it will place the
OB. 1 04
new one above the selected one.
g = adds a glass cube.
<shift> = for multiple selections.
s = Save, a file called "newfile.iv " will be created in the directory that
emacs was launched from.
Selecting the "hand" icon will allow one to move and pan through the virtual space. The movements have been constrained to "walking" in a plane,
one may "pan" up but this is a conscious decision. There is no "flying", I
hope.
OB. 105
.iv FILES AND ALTERATIONS
This section is necessary to get PHEMENOarch
to readin its own .iv files and .iv files from other
programs like Formez, architrinon, Alias,
Softimage, ProEngineer, Catia, and autocad. It
also will explain how to get other programs (like
ivview) to read the .iv files produced with
PHEMENOarch.
PHEMENOarch writes out .iv files in a very particular manner. This is mainly due to the way the
SoDB and Scene graphs had to be written to
allow for the manipulation of primitives in free
space.
Here is an example of the Open Inventor format
.iv file that is generated directly from the scheme
interpreter:
#lnventor V2.1 ascii
Separator {
PointLight {
TRUE
on
location
20 20 101
}
Selection {
Oc. 106
policy SHIFT
EventCallback {
I
Separator {
Transform {
-1.42036 1.72645 2.48729e-08
translation
rotation
0.892297 -0.429779 0.13819 0.691726
scaleFactor 0.832692 0.832692 0.832691
scaleOrientation
-0.647332 0.248573 -0.720537 0.663988
i
Material {
diffuseColor
11 1
Texture2 {
filename
"img.rgb"
Cube {
Separator {
TabBoxManip {
translation
0.267307 0.8781 8 0
Material {
diffuseColor
1 11
}
Texture2 {
filename
"img.rgb"
Cube {
}
Separator {
Transform {
translation
rotation
scaleFactor
1.54463 1.15541 1.14504
0 01 0
1.13823 1.10481 1
I
Material {
diffuseColor
0.5 0.5 0.5
Oc. 107
Texture2 {
filename
"Conc.rgb"
Cube {
}
Separator {
Transform {
translation
scaleFactor
-0.479954 0.908418 1.57299
1.88592 0.27275 1
}
Material {
diffuseColor
specularColor
shininess
transparency
00.20
0 1 0.2
0.1
0.6
Texture2 {
filename
"img.rgb"
Cube {
}
Note that if the file was saved while a primitive
was selected. The Manipulator will still be
attached to the oblect. This needs to be altered
to be read into other programs, including itself.
The easiest way to avoid this, is to make sure that
nothing is selected before saving. It can still be
altered afterward through a text editor like jot or
Emacs.
Oc. 108
In order for PHEMENOarch to be able to read in
an .iv file (regardless of the source of the .iv), it
must first be altered to look like this:
#lnventor V2.1 ascii
Separator {
Transform {
translation
-1.42036 1.72645 2.48729e-08
rotation
0.892297 -0.429779 0.13819 0.691726
scaleFactor
0.832692 0.832692 0.832691
scaleOrientation
-0.647332 0.248573 -0.720537 0.663988
}
Material{
diffuseColor
11 1
}
Texture2
filename
"img.rgb"
}
Cube {
}
}
Separator {
Transform {
translation
0.267307 0.8781 8 0
}
Material {
diffuseColor
1 11
}
Texture2
{
Oc. 109
filename
"1img.rgb"
Cube {
Separator {
Transform {
translation
1.54463 1.15541 1.14504
rotation
00 1 0
1.13823 1.1048] 1
scaleFactor
}
Material {
diffuseColor
0.5 0.5 0.5
Texture2 {
filename
"Conc. rgb"
Cube {
}
Separator {
Transform {
translation
-0.479954 0.908418 1.57299
scaleFactor
1.88592 0.27275 1
Material {
diffuseColor 0 0.2 0
specularColor 0 1 0.2
0.1
shininess
transparency 0.6
Oc. I 1 0
-
2MML1___
Texture2
{
filename
"img.rgb"
}
Cube {
}
}
The SoSeperator of the .iv file must be placed at
the top of the Scene graph. The transforms must
be placed right underneath the separator. A failure to do so will result in the model not being
manipulatable within PHEMENOarch, you will be
able to add new geometry and this new geometry will be manipulatable but the imported geometry will NOT.
In order to read the resultant PHEMENOarch .iv
files in other programs and viewers, one must
make a few changes. Fore instance in writing for
"ivview" a standard Open Inventor viewer for the
SGI, one must first find out the correct version of
Inventor ascii that the viewer is looking for. In this
case the current "ivview" wanted version 2.0 and
not 2.1 which PHEMENOarch was writing.
Oc. I I I
Here is an example of an .iv file readable by "ivview";
#Inventor V2.0 ascii
Separator {
PointLight {
on
TRUE
location
2020 101
Separator {
Separator {
Transform {
translation
-1.42036 1.72645
2 .4 8 729e-08
0.892297 -0.429779 0.13819 0.691726
scaleFactor 0.832692 0.832692 0.832691
-0.647332 0.248573 -0.720537 0.663988
scaleOrientation
rotation
Material {
diffuseColor
11 1
Texture2 {
filename
"img.rgb"
Cube {
}
}
Separator {
Transform {
translation
0.267307 0.87818 0
i
Material {
diffuseColor
111
Texture2 {
filename
"img .rgb"
Cube {
I
Oc. I
2
Separator {
Transform {
translation
rotation
scaleFactor
Material {
diffuseColor
1.54463 1.15541 1.14504
00 1 0
1.13823 1.10481
0.5 0.5 0.5
}
Texture2
filename
"Conc.rgb"
Cube {
}
}
Separator {
Transform {
translation
scaleFactor
-0.479954 0.90841 8 1.57299
1.88592 0.27275 1
Material {
diffuseColor 0 0.2 0
speculairColor 0 1 0.2
0.1
shininess
transparency 0.6
}
Texture2 {
filename
"img.rgb"
Cube {
}
Oc.I 13
The selection callback node that PHEMENOarch produces was simply replaced with another separator. I.E.
Selection
{
policy SHIFT
EventCallback {
Changed to just:
Separator
{
With these nuances in mind, PHEMENOarch will work
with any .iv files that can be created with any other program.
Oc.I 14
BIBLIOGRAPHY
Ching, Francis, Arch itecture: Form, Space & Order, NewYork: Van
Nostrand Reinhold, 1979
Critchlow, Keith, Order in Space, New York: Viking Press, 1965
Foley, J.Introduction to Computer Graphics, New York: Addison-Wesley,
1994
GABB, The Great American Bathroom Book, Salt Lake: Compact Classics,
1991
Gasson, Peter C., Geometry of Spatial Forms, New York: Ellis Horwood
Limited, 1983
Haber, R.N. The psycology of Visual Perception. New York: Holt, Rinehart
& Winston, 1980
Hanks, David A., Frank Lloyd Wright, New York: E. P.Dutton, 1989
Kerlow, J. The Art of 3-D Computer Animation and Imaging, NewYork: Van
Nostrand Reinhold, 1996
Kubovy, Michael, The Psychology of Perspective and Renaissance Art,
Cambridge: Cambridge Press , 1986
BIB.I 1 5
A
BIBLIOGRAPHY ||
Leonardo da Vinci. The note books of Leonardo da Vinci. Ed j.p. Richter.
London: Phaidon, 1970
Mitchell, William
J. and McCullough, Malcolm , Digital Design Media.
NewYork: Van Nostrand Reinhold, 1995
Morehead, James, A Handbook of Perspective Drawing, Florida: H & W.
B. Drew Co. 1941
Pedoe, Dan, Geometry and the Visual Arts, New York: Dover Publications,
1976
Solso, Robert, Cognition and the Visual Arts. Cambridge: MIT Press 1994
Smith, P.C. , and Smith, O.W. Ball-throwing responces to photographically
portrayed targets. lournal of Experamental psycology, 1961, 62, 223-33
Thompson, Kristin and Bordwell, David, Film History, an Introduction, New
York: McGraw-Hill, 1994
Wernecke, Josie ,The Inventor Mentor. Reading, Mass.: Addison-Wesley,
1994
Woods,Lebbeus, War and Architecture, Princeton: Princeton Architectural
Press, 1993.
BIB.I 16